From owner-linux-xfs@oss.sgi.com Sat Dec 1 01:42:58 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB19gwm23104 for linux-xfs-outgoing; Sat, 1 Dec 2001 01:42:58 -0800 Received: from smtpzilla5.xs4all.nl (smtpzilla5.xs4all.nl [194.109.127.141]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB19gro23082 for ; Sat, 1 Dec 2001 01:42:54 -0800 Received: from auto-nb1.xs4all.nl (coltex.xs4all.nl [213.84.127.168]) by smtpzilla5.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB18ggGG005710; Sat, 1 Dec 2001 09:42:48 +0100 (CET) Message-Id: <4.3.2.7.2.20011201093658.0373bf48@pop.xs4all.nl> X-Sender: knuffie@pop.xs4all.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Sat, 01 Dec 2001 09:39:42 +0100 To: "John Stevens" , From: Seth Mos Subject: Re: Problem with XFS 1.0.2a In-Reply-To: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 22:43 30-11-2001 -0800, John Stevens wrote: >I downloaded the ISO image, and did a clean install. The 3com (3C985) Gig E >card doesn't work. If I do a standard install using just the Redhat 7.2 CDs >the card works fine. The System is an Intel SBT2 Motherboard, with 2 PCI >cards 3C905 and an Adaptec 39160. Any thoughts? Hmm, would that be the Copper Gigabit Ethernet card? modprobe bc5200 It should be autodetected. The card has been around for a while now. If you know what module it uses you could switch to the second console ctrl-alt-F2 and do a modprobe. Cheers -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Sat Dec 1 12:38:35 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB1KcZw07023 for linux-xfs-outgoing; Sat, 1 Dec 2001 12:38:35 -0800 Received: from mailout01.sul.t-online.de (mailout01.sul.t-online.com [194.25.134.80]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB1KcUo07000 for ; Sat, 1 Dec 2001 12:38:30 -0800 Received: from fwd06.sul.t-online.de by mailout01.sul.t-online.de with smtp id 16AFxV-0008QH-04; Sat, 01 Dec 2001 20:37:57 +0100 Received: from tower (340024412816-0001@[217.230.8.33]) by fwd06.sul.t-online.com with esmtp id 16AFxD-1LpiRkC; Sat, 1 Dec 2001 20:37:39 +0100 Content-Type: text/plain; charset="iso-8859-1" From: Hasch@t-online.de (Juergen Hasch) To: ivanr@sgi.com Subject: Re: Again: xfsrestore assertion failure Date: Sat, 1 Dec 2001 20:37:27 +0100 X-Mailer: KMail [version 1.3.7] Cc: linux-xfs@oss.sgi.com References: <169t5H-0FoJM0C@fwd03.sul.t-online.com> In-Reply-To: <169t5H-0FoJM0C@fwd03.sul.t-online.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-ID: <16AFxD-1LpiRkC@fwd06.sul.t-online.com> X-Sender: 340024412816-0001@t-dialin.net Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Am Friday 30 November 2001 20:12 schrieb Juergen Hasch: > In line 1508 of drive_scsitape.c I find the following values: > first_mark_offset = 206e460000000000 > file_offset = 400000 > tape_recsz = 809efa800100000 Ivan, things go wrong in xfsdump. When do_set_mark() sets first_mark_offset in the rec_hdr_t struct it doesn't endian convert it. It is also not endian converted later The following patch solves this, however you could also do the endian conversion later in do_write(). ...Juergen --- cmd/xfsdump/common/drive_scsitape.c.orig Fri Nov 30 19:43:48 2001 +++ cmd/xfsdump/common/drive_scsitape.c Sat Dec 1 20:25:35 2001 @@ -1940,7 +1940,7 @@ rechdrp = ( rec_hdr_t * )contextp->dc_recp; if ( rechdrp->first_mark_offset == -1LL ) { ASSERT( nextoff != -1LL ); - rechdrp->first_mark_offset = nextoff; + rechdrp->first_mark_offset = INT_GET(nextoff,ARCH_CONVERT); } /* put the mark on the tail of the queue. From owner-linux-xfs@oss.sgi.com Sat Dec 1 17:16:24 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB21GOT11192 for linux-xfs-outgoing; Sat, 1 Dec 2001 17:16:24 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB21GGo11166 for ; Sat, 1 Dec 2001 17:16:16 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via SMTP id BAA929807 for ; Sun, 2 Dec 2001 01:15:54 +0100 (CET) mail_from (ivanr@sgi.com) From: ivanr@sgi.com Received: from omen.melbourne.sgi.com (omen.melbourne.sgi.com [134.14.55.139]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id LAA02756; Sun, 2 Dec 2001 11:14:18 +1100 Received: from localhost (ivanr@localhost) by omen.melbourne.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id LAA55214; Sun, 2 Dec 2001 11:14:17 +1100 (EST) X-Authentication-Warning: omen.melbourne.sgi.com: ivanr owned process doing -bs Date: Sun, 2 Dec 2001 11:14:17 +1100 X-X-Sender: ivanr@omen.melbourne.sgi.com To: Juergen Hasch cc: linux-xfs@oss.sgi.com Subject: Re: Again: xfsrestore assertion failure In-Reply-To: <16AFxD-1LpiRkC@fwd06.sul.t-online.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Sat, 1 Dec 2001, Juergen Hasch wrote: > Am Friday 30 November 2001 20:12 schrieb Juergen Hasch: > > > In line 1508 of drive_scsitape.c I find the following values: > > first_mark_offset = 206e460000000000 > > file_offset = 400000 > > tape_recsz = 809efa800100000 > > Ivan, things go wrong in xfsdump. > When do_set_mark() sets first_mark_offset in the rec_hdr_t struct > it doesn't endian convert it. It is also not endian converted later > The following patch solves this, however you could also do the > endian conversion later in do_write(). > ...Juergen Thanks very much for your investigation! Tim or I will review this patch in the next day or two. Ivan > > --- cmd/xfsdump/common/drive_scsitape.c.orig Fri Nov 30 19:43:48 2001 > +++ cmd/xfsdump/common/drive_scsitape.c Sat Dec 1 20:25:35 2001 > @@ -1940,7 +1940,7 @@ > rechdrp = ( rec_hdr_t * )contextp->dc_recp; > if ( rechdrp->first_mark_offset == -1LL ) { > ASSERT( nextoff != -1LL ); > - rechdrp->first_mark_offset = nextoff; > + rechdrp->first_mark_offset = INT_GET(nextoff,ARCH_CONVERT); > } > > /* put the mark on the tail of the queue. > > -- Ivan Rayner ivanr@sgi.com From owner-linux-xfs@oss.sgi.com Sat Dec 1 17:50:33 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB21oXj11835 for linux-xfs-outgoing; Sat, 1 Dec 2001 17:50:33 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB21oTo11803 for ; Sat, 1 Dec 2001 17:50:29 -0800 Received: from relay1.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id QAA06968 for ; Sat, 1 Dec 2001 16:50:27 -0800 (PST) mail_from (sandeen@sgi.com) Received: from chuckle.americas.sgi.com (sandeen@chuckle.americas.sgi.com [128.162.211.44]) by relay1.corp.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id QAA92525; Sat, 1 Dec 2001 16:49:57 -0800 (PST) Date: Sat, 1 Dec 2001 18:49:56 -0600 (CST) From: Eric Sandeen X-X-Sender: To: John Stevens cc: Subject: Re: Problem with XFS 1.0.2a In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Doesn't work during the install, or doesn't work on the installed, running system? The XFS installer runs under the same basic 2.4.7 kernel as the stock Red Hat installer, but puts the updated 2.4.9 kernels on your system. Does the card work on the updated 2.4.9 kernel from Red Hat? -Eric On Fri, 30 Nov 2001, John Stevens wrote: > I downloaded the ISO image, and did a clean install. The 3com (3C985) Gig E > card doesn't work. If I do a standard install using just the Redhat 7.2 CDs > the card works fine. The System is an Intel SBT2 Motherboard, with 2 PCI > cards 3C905 and an Adaptec 39160. Any thoughts? > > John Stevens > Engineer / Sysadmin > Encore Hollywood > From owner-linux-xfs@oss.sgi.com Sun Dec 2 07:20:40 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB2FKeb00361 for linux-xfs-outgoing; Sun, 2 Dec 2001 07:20:40 -0800 Received: from mail.1ar.org ([202.88.158.15]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB2FK8o00330 for ; Sun, 2 Dec 2001 07:20:09 -0800 Received: from localhost (mail.1ar.org [127.0.0.1]) by mail.1ar.org (Postfix) with ESMTP id 00AC41007BFC; Sun, 2 Dec 2001 19:49:54 +0530 (IST) Date: Sun, 2 Dec 2001 19:49:54 +0530 (IST) From: Ajay Ramaswamy X-X-Sender: To: Cc: Subject: Problems with XFS 1.0.2a anaconda In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hello, I have just stated playing around with the installer and making some modifications to it when I noticed some problems 1. Why is the installer not adding xfsdump and dependencies to the Utilities group? Look at the diffs below from XFS 1.0.1 for RedHat 7.1 --- comps.rh71 Sun Dec 2 19:07:40 2001 +++ comps.sgi71 Sun Dec 2 19:06:48 2001 @@ -134,6 +134,7 @@ vixie-cron which words + xfsprogs zlib (lang ja and arch i386): kon2 (lang ja and arch i386): kon2-fonts @@ -843,6 +844,7 @@ patch rcs e2fsprogs-devel + xfsprogs-devel binutils gcc sparc: gcc-sparc32 @@ -975,6 +977,8 @@ 0 Kernel Development { @ Development kernel-source + !ia64: compat-egcs + !ia64: compat-glibc } 0 Utilities { @@ -991,6 +995,10 @@ bc rmt dump + xfsdump + attr + acl + dmapi efax make lsof and from XFS 1.0.2a and RedHat 7.2 --- disc1/RedHat/base/comps Mon Sep 10 06:22:59 2001 +++ xfs/RedHat/base/comps Fri Oct 26 07:44:25 2001 @@ -121,6 +121,8 @@ redhat-logos redhat-release (arch !s390 and arch !s390x): reiserfs-utils + (arch !s390 and arch !s390x): xfsprogs + lvm-tools rootfiles rpm s390: s390utils 2. In the anaconda rpm --- anaconda-7.2-rh/scripts/mk-images.i386 Thu Sep 6 03:55:45 2001 +++ anaconda-7.2-sgi/scripts/mk-images.i386 Sun Dec 2 18:20:24 2001 @@ -80,7 +80,7 @@ IDEMODS="" SCSIMODS="sd_mod sr_mod" -SECSTAGE="agpgart msdos vfat raid0 raid1 raid5 ext3 reiserfs $IDEMODS $SCSIMODS $LATEUSBMODS" +SECSTAGE="agpgart msdos vfat raid0 raid1 raid5 ext3 reiserfs xfs $IDEMODS $SCSIMODS $LATEUSBMODS" COMMONMODULES="vfat $USBMODS" LOCALMODULES="$COMMONMODULES BusLogic aic7xxx megaraid ncr53c8xx I have found that using "pagebuf xfs_support xfs" instead of just "xfs" is more reliable and always genarates the correct iso for me otherwise using the 2.4.9-13 boot kernel does not work for me. 3. In creating the install image for the CD --- anaconda-7.2-rh/scripts/upd-instroot Wed Aug 22 22:03:26 2001 +++ anaconda-7.2-sgi/scripts/upd-instroot Sun Dec 2 18:20:24 2001 @@ -95,7 +95,7 @@ ash kon2 e2fsprogs util-linux raidtools popt mount procps rpm XFree86 Xconfigurator anaconda anaconda-runtime kudzu-devel kudzu db3 rpm-python bzip2 bzip2-libs dosfstools pciutils - reiserfs-utils parted busybox-anaconda" + reiserfs-utils xfsprogs xfsdump parted busybox-anaconda" if [ $ARCH = i386 ]; then PACKAGES="$PACKAGES kernel-pcmcia-cs lilo" @@ -121,7 +121,8 @@ XFree86-cyrillic-fonts XFree86 Xconfigurator pygnome pygtk pygtk-libglade pygnome-libglade pygnome-applet pygnome-capplet gdk-pixbuf XFree86-KOI8-R - XFree86-KOI8-R-75dpi-fonts pciutils pam reiserfs-utils" + XFree86-KOI8-R-75dpi-fonts pciutils pam reiserfs-utils xfsprogs + xfsdump" if [ -z "$NEEDGR" ]; then PACKAGESGR="bash" @@ -202,11 +203,17 @@ sbin/mkfs.msdos sbin/mkfs.vfat sbin/mkreiserfs +sbin/mkfs.xfs sbin/mkraid sbin/mkswap sbin/parted sbin/probe sbin/tune2fs +sbin/xfsdump +sbin/xfsrestore +sbin/xfs_repair +usr/sbin/xfs_db +usr/sbin/xfs_check usr/X11R6/lib/X11/Cards usr/X11R6/share/Xconfigurator/MonitorsDB usr/bin/python* @@ -296,6 +303,7 @@ sbin/fsck sbin/fsck.ext2 sbin/fsck.ext3 +sbin/fsck.xfs sbin/ifconfig sbin/insmod sbin/lsmod as I can see it xfsdump will not run since the necessary libraries are nowhere to be found # ldd /sbin/xfsdump libattr.so.0 => /lib/libattr.so.0 (0x40030000) libdm.so.0 => /lib/libdm.so.0 (0x40033000) libhandle.so.0 => /lib/libhandle.so.0 (0x40039000) libc.so.6 => /lib/libc.so.6 (0x4003c000) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000) # libhandle is from xfsprogs, libdm is from dmapi & libattr is from attr 4. When I use the XFS 1.0.2 cd in rescue mode it complains that it cannot find the filesystems that are specified by LABEL= tags in fstab. I see that the label is assigned by + def labelDevice(self, entry, chroot): + devicePath = entry.device.setupDevice(chroot) + label = labelFactory.createLabel(entry.mountpoint, self.maxLabelChars) + db_cmd = "label " + label + rc = iutil.execWithRedirect("/usr/sbin/xfs_db", + ["xfs_db", "-x", "-c", db_cmd, devicePath],+ stdout = "/dev/tty5", + stderr = "/dev/tty5") + if rc: + raise SystemError + entry.setLabel(label) + in dispatch.py, but when I do man xfs_db I cannot find any way to set the label with command label. Thank you regards Ajay From owner-linux-xfs@oss.sgi.com Sun Dec 2 09:34:12 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB2HYCA03060 for linux-xfs-outgoing; Sun, 2 Dec 2001 09:34:12 -0800 Received: from softhome.net (jive.SoftHome.net [66.54.152.27]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB2HYAo03038 for ; Sun, 2 Dec 2001 09:34:10 -0800 Received: from softhome.net ([217.127.11.76]) (AUTH: PLAIN colario@softhome.net) by softhome.net with esmtp; Sun, 02 Dec 2001 09:23:34 -0700 Message-ID: <3C0A5892.7040101@softhome.net> Date: Sun, 02 Dec 2001 16:36:34 +0000 From: Ayose User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.6) Gecko/20011127 X-Accept-Language: es, es-es MIME-Version: 1.0 To: linux-xfs@oss.sgi.com Subject: XFS In 2.5 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk In 2.5 will be added XFS? I have seen a lot of discussions about that inclussion, but I can't get a definitive answer :( From owner-linux-xfs@oss.sgi.com Sun Dec 2 10:34:55 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB2IYtr04099 for linux-xfs-outgoing; Sun, 2 Dec 2001 10:34:55 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB2IYoo04076 for ; Sun, 2 Dec 2001 10:34:50 -0800 Received: from relay1.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id JAA04022 for ; Sun, 2 Dec 2001 09:34:48 -0800 (PST) mail_from (sandeen@sgi.com) Received: from chuckle.americas.sgi.com (sandeen@chuckle.americas.sgi.com [128.162.211.44]) by relay1.corp.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id JAA79140; Sun, 2 Dec 2001 09:34:17 -0800 (PST) Date: Sun, 2 Dec 2001 11:34:16 -0600 (CST) From: Eric Sandeen X-X-Sender: To: Ajay Ramaswamy cc: , Subject: Re: Problems with XFS 1.0.2a anaconda In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi Ajay, I'll let mkp respond to most of this ;-) but: On Sun, 2 Dec 2001, Ajay Ramaswamy wrote: > 4. When I use the XFS 1.0.2 cd in rescue mode it complains that it cannot > find the filesystems that are specified by LABEL= tags in fstab. I see > that the label is assigned by > > + def labelDevice(self, entry, chroot): [etc...] > in dispatch.py, but when I do man xfs_db I cannot find any way to set the > label with command label. Well, this is a bit tricky. If you look at the man page for xfs_admin, you'll see the label command. Then look at /usr/sbin/xfs_admin, and you'll see that it's just a script that calls xfs_db. Perhaps the labeling info should be in the xfs_db man page as well. I caled xfs_db here instead of xfs_admin just so that we wouldn't have to add xfs_admin to the install environment. However, I'm not sure why rescue mode can't find filesystems with LABELs in fstab, I'm pretty sure this works over here. And as long as the LABELs have been made correctly, it's then up to "mount" to decipher labels on devices, I think. -Eric From owner-linux-xfs@oss.sgi.com Sun Dec 2 10:42:21 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB2IgLr04318 for linux-xfs-outgoing; Sun, 2 Dec 2001 10:42:21 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB2IgHo04296 for ; Sun, 2 Dec 2001 10:42:17 -0800 Received: from relay1.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id JAA07434 for ; Sun, 2 Dec 2001 09:42:15 -0800 (PST) mail_from (sandeen@sgi.com) Received: from chuckle.americas.sgi.com (sandeen@chuckle.americas.sgi.com [128.162.211.44]) by relay1.corp.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id JAA69731; Sun, 2 Dec 2001 09:41:45 -0800 (PST) Date: Sun, 2 Dec 2001 11:41:44 -0600 (CST) From: Eric Sandeen X-X-Sender: To: Ayose cc: Subject: Re: XFS In 2.5 In-Reply-To: <3C0A5892.7040101@softhome.net> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Short answer: It's up to Linus, ask him. :) Long answer: Getting XFS into 2.5 will take work, both technical and political. Technical - The 2.5 block layer has changed quite a bit already (as of 2.5.1-pre2, I believe) and at least the pagebuf interface will need to be modified to handle this. Steve has started looking at the block layer modifications. Political - (maybe this is technical as well) - previous discussions on LKML have shown that some of the Linux hackers would like to see the XFS <-> Linux interface changed. You can check the archives for this discussion, but this is the sort of thing that will have to be worked out for XFS to get into 2.5, I think. We do have a 2.5 tree internally, and we're working on it. It will be exported via CVS sometime soon. -Eric On Sun, 2 Dec 2001, Ayose wrote: > In 2.5 will be added XFS? > > I have seen a lot of discussions about that inclussion, but I can't get > a definitive answer :( > From owner-linux-xfs@oss.sgi.com Sun Dec 2 11:23:37 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB2JNbl04944 for linux-xfs-outgoing; Sun, 2 Dec 2001 11:23:37 -0800 Received: from chaos.egr.duke.edu (chaos.egr.duke.edu [152.3.195.82]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB2JNXo04922 for ; Sun, 2 Dec 2001 11:23:34 -0800 Received: from localhost (jlb@localhost) by chaos.egr.duke.edu (8.11.6/8.11.6) with ESMTP id fB2INUj07327; Sun, 2 Dec 2001 13:23:30 -0500 X-Authentication-Warning: chaos.egr.duke.edu: jlb owned process doing -bs Date: Sun, 2 Dec 2001 13:23:30 -0500 (EST) From: Joshua Baker-LePain X-X-Sender: To: Eric Sandeen cc: Linux xfs mailing list Subject: Re: XFS In 2.5 In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Sun, 2 Dec 2001 at 11:41am, Eric Sandeen wrote > We do have a 2.5 tree internally, and we're working on it. It will be > exported via CVS sometime soon. > Now that 2.5 has opened, and assuming that you have CVS trees based on both 2.4.x and 2.5.x, what will be the "purpose" of the trees? Will the 2.4.x tree still be the "main" tree, where bugfixes and feature enhancements go, to then be ported the the 2.5.x tree? Given that, the main purpose of the 2.5.x tree would be to keep up with the changes of 2.5? Or will most work be done on the 2.5.x tree, with bug fixes and enhancements back ported to 2.4.x? Just curious, really... Thanks! -- Joshua Baker-LePain Department of Biomedical Engineering Duke University From owner-linux-xfs@oss.sgi.com Sun Dec 2 11:51:35 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB2JpZp05495 for linux-xfs-outgoing; Sun, 2 Dec 2001 11:51:35 -0800 Received: from etoile.infra.idealx.com (aboukir-101-1-1-jtournier.adsl.nerim.net [62.4.18.139]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB2JpVo05472 for ; Sun, 2 Dec 2001 11:51:31 -0800 Received: (qmail 793 invoked by uid 500); 2 Dec 2001 18:51:21 -0000 From: "Jérôme Tournier" Date: Sun, 2 Dec 2001 19:51:21 +0100 To: Ajay Ramaswamy Cc: linux-xfs@oss.sgi.com Subject: Re: anaconda-runtime-7.2-11XFS problem Message-ID: <20011202195120.A776@etoile.tux.iallanis.com> References: <20011130095142.A1943@etoile.magasin1.auchan.com> <1007138106.16790.7.camel@stout.americas.sgi.com> <20011130181216.B1943@etoile.magasin1.auchan.com> <1007187459.3c08760325e10@mail.1ar.org> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <1007187459.3c08760325e10@mail.1ar.org> User-Agent: Mutt/1.3.23i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Le Sat, Dec 01, 2001 at 11:47:39AM +0500, Ajay Ramaswamy à écrit: # I had the same problem and I fixed it by editing /usr/lib/anaconda-runtime/mk- # images.i386, here there is a line to load all the fs modules ie msdos vfat ext3 # reiser and xfs, I changed it to include pagebuf and xfs_support Well, i tried this but i have the dependancies problem. What is very strange is that sometimes, when i run the buildinstall script, all the xfs modules are loaded . Sometimes it work fine but most of the times, it's not good !! -- Jérôme From owner-linux-xfs@oss.sgi.com Sun Dec 2 15:05:30 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB2N5Ur09398 for linux-xfs-outgoing; Sun, 2 Dec 2001 15:05:30 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB2N5Po09374 for ; Sun, 2 Dec 2001 15:05:25 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via SMTP id OAA13637 for ; Sun, 2 Dec 2001 14:05:12 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id JAA06785; Mon, 3 Dec 2001 09:04:06 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id JAA41614; Mon, 3 Dec 2001 09:04:05 +1100 (AEDT) Date: Mon, 3 Dec 2001 09:04:05 +1100 From: Nathan Scott To: Ajay Ramaswamy Cc: mkp@linuxcare.com, linux-xfs@oss.sgi.com Subject: Re: Problems with XFS 1.0.2a anaconda Message-ID: <20011203090405.B39338@wobbly.melbourne.sgi.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from ajayr@krithika.net on Sun, Dec 02, 2001 at 07:49:54PM +0530 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Sun, Dec 02, 2001 at 07:49:54PM +0530, Ajay Ramaswamy wrote: > Hello, > I have just stated playing around with the installer and making some > modifications to it when I noticed some problems > ... [I'll leave the other questions to Eric/Martin as they know much more about the Redhat installer than I do.] > 4. When I use the XFS 1.0.2 cd in rescue mode it complains that it cannot > find the filesystems that are specified by LABEL= tags in fstab. Do you know what complains? Is it mount? fsck? or something else? The code to support XFS labels has been in mount for a long time, so I would not expect it to be the problem. fsck was only updated more recently (2/3 months ago) so this might be the problem (might need a newer version of e2fsprogs on the rescue CD?). > I see that the label is assigned by > > + def labelDevice(self, entry, chroot): > + devicePath = entry.device.setupDevice(chroot) > + label = labelFactory.createLabel(entry.mountpoint, > self.maxLabelChars) > + db_cmd = "label " + label > + rc = iutil.execWithRedirect("/usr/sbin/xfs_db", > + ["xfs_db", "-x", "-c", db_cmd, > devicePath],+ stdout = "/dev/tty5", > + stderr = "/dev/tty5") > + if rc: > + raise SystemError > + entry.setLabel(label) > + > > in dispatch.py, but when I do man xfs_db I cannot find any way to set the > label with command label. > I'll add something to the man page today. thanks. -- Nathan From owner-linux-xfs@oss.sgi.com Sun Dec 2 16:12:38 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB30Ccn10967 for linux-xfs-outgoing; Sun, 2 Dec 2001 16:12:38 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB30CXo10944 for ; Sun, 2 Dec 2001 16:12:33 -0800 Received: from relay1.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id PAA22799 for ; Sun, 2 Dec 2001 15:12:22 -0800 (PST) mail_from (sandeen@sgi.com) Received: from chuckle.americas.sgi.com (sandeen@chuckle.americas.sgi.com [128.162.211.44]) by relay1.corp.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id PAA68035; Sun, 2 Dec 2001 15:12:01 -0800 (PST) Date: Sun, 2 Dec 2001 17:12:01 -0600 (CST) From: Eric Sandeen X-X-Sender: To: Joshua Baker-LePain cc: Linux xfs mailing list Subject: Re: XFS In 2.5 In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi Joshua - That's still under discussion, really, but since 2.4.x will have more users, and probably more bug reports, I would guess that feature & bugfix work will happen primarily in 2.4.x, pushed over to 2.5.x as appropriate. 2.5 will probably be working on the block layer interface, etc, and stay busy enough with that activity. That's just my hunch, subject to review and change. :) -Eric On Sun, 2 Dec 2001, Joshua Baker-LePain wrote: > On Sun, 2 Dec 2001 at 11:41am, Eric Sandeen wrote > > > We do have a 2.5 tree internally, and we're working on it. It will be > > exported via CVS sometime soon. > > > Now that 2.5 has opened, and assuming that you have CVS trees based on > both 2.4.x and 2.5.x, what will be the "purpose" of the trees? Will the > 2.4.x tree still be the "main" tree, where bugfixes and feature > enhancements go, to then be ported the the 2.5.x tree? Given that, the > main purpose of the 2.5.x tree would be to keep up with the changes of > 2.5? Or will most work be done on the 2.5.x tree, with bug fixes and > enhancements back ported to 2.4.x? > > Just curious, really... Thanks! > > From owner-linux-xfs@oss.sgi.com Sun Dec 2 17:06:00 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB3160W12254 for linux-xfs-outgoing; Sun, 2 Dec 2001 17:06:00 -0800 Received: from starship.berlin (dsl-213-023-038-056.arcor-ip.net [213.23.38.56]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB315to12231 for ; Sun, 2 Dec 2001 17:05:56 -0800 Received: from daniel by starship.berlin with local (Exim 3.32 #1 (Debian)) id 16Agdh-0000BS-00; Mon, 03 Dec 2001 01:07:17 +0100 Content-Type: text/plain; charset="iso-8859-1" From: Daniel Phillips To: Nathan Scott , Alexander Viro , Andi Kleen , Andreas Gruenbacher , Linus Torvalds , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [RFC][PATCH] VFS interface for extended attributes Date: Mon, 3 Dec 2001 01:07:13 +0100 X-Mailer: KMail [version 1.3.2] References: <20011114230134.A5739@lynx.no> <20011116101800.A632931@wobbly.melbourne.sgi.com> In-Reply-To: <20011116101800.A632931@wobbly.melbourne.sgi.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, sorry for jumping into this a little late, but... On November 16, 2001 12:18 am, Nathan Scott wrote: > > What is the distinction between "set" and "replace" or "set" and "create"? > > +#define EA_CREATE 0x0001 /* Set the value: fail if attr already exists */ > +#define EA_REPLACE 0x0002 /* Set the value: fail if attr does not exist */ > > Whereas "set" is simply set the named attribute value, creating the > attribute if need be, replacing the value if the attribute exists, > and then return success. What is the purpose of these distinctions? Does anyone rely on them? Do such distinctions exist in an existing implementation? Thanks. Daniel From owner-linux-xfs@oss.sgi.com Sun Dec 2 17:55:31 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB31tVt13044 for linux-xfs-outgoing; Sun, 2 Dec 2001 17:55:31 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB31tPo13015 for ; Sun, 2 Dec 2001 17:55:25 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via SMTP id QAA06989 for ; Sun, 2 Dec 2001 16:55:23 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id LAA07737; Mon, 3 Dec 2001 11:54:02 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id LAA41797; Mon, 3 Dec 2001 11:54:01 +1100 (AEDT) Date: Mon, 3 Dec 2001 11:54:01 +1100 From: Nathan Scott To: Daniel Phillips Cc: Alexander Viro , Andi Kleen , Andreas Gruenbacher , Linus Torvalds , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [RFC][PATCH] VFS interface for extended attributes Message-ID: <20011203115400.F39338@wobbly.melbourne.sgi.com> References: <20011114230134.A5739@lynx.no> <20011116101800.A632931@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from phillips@bonn-fries.net on Mon, Dec 03, 2001 at 01:07:13AM +0100 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi Daniel, On Mon, Dec 03, 2001 at 01:07:13AM +0100, Daniel Phillips wrote: > Hi, sorry for jumping into this a little late, but... > No problem. BTW, we have reworked the interfaces once more and will send out the latest revision in the next couple of days - it does away with commands and flags completely, except for this one instance of flags in the set operation... > On November 16, 2001 12:18 am, Nathan Scott wrote: > > > What is the distinction between "set" and "replace" or "set" and "create"? > > > > +#define EA_CREATE 0x0001 /* Set the value: fail if attr already exists */ > > +#define EA_REPLACE 0x0002 /* Set the value: fail if attr does not exist */ > > > > Whereas "set" is simply set the named attribute value, creating the > > attribute if need be, replacing the value if the attribute exists, > > and then return success. > > What is the purpose of these distinctions? Does anyone rely on them? Do such > distinctions exist in an existing implementation? > The purpose is to provide user tools with more control over the creation or updating of an attribute and its value. I don't think the replace flag is very widely used, but I have seen the create flag used in places - eg. the XFS fsr tool uses that flag. The IRIX extended attribute interfaces provide these flags - Andreas has also examined the implementations, man pages, etc, of several other operating systems, so he'll be able to tell us if any others provide this sort of thing too. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Sun Dec 2 18:01:06 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB3216i13361 for linux-xfs-outgoing; Sun, 2 Dec 2001 18:01:06 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB3213o13336 for ; Sun, 2 Dec 2001 18:01:03 -0800 Received: from snort.melbourne.sgi.com (snort.melbourne.sgi.com [134.14.55.149]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id CAA1148285 for ; Mon, 3 Dec 2001 02:00:59 +0100 (CET) mail_from (nathans@snort.melbourne.sgi.com) Received: (from nathans@localhost) by snort.melbourne.sgi.com (SGI-8.9.3/8.9.3) id LAA13656 for linux-xfs@oss.sgi.com; Mon, 3 Dec 2001 11:59:28 +1100 (EST) Date: Mon, 3 Dec 2001 11:59:28 +1100 (EST) From: Nathan Scott Message-Id: <200112030059.LAA13656@snort.melbourne.sgi.com> To: linux-xfs@oss.sgi.com Subject: TAKE - xfs_db man page Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Sun Dec 2 16:58:29 PST 2001 Workarea: snort.melbourne.sgi.com:/home/nathans/linux-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107623a cmd/xfsprogs/man/man8/xfs_db.8 - 1.4 - Document the "label" and "uuid" commands. From owner-linux-xfs@oss.sgi.com Sun Dec 2 23:11:01 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB37B1l26181 for linux-xfs-outgoing; Sun, 2 Dec 2001 23:11:01 -0800 Received: from zok.sgi.com (zok.sgi.com [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB37Awo26155 for ; Sun, 2 Dec 2001 23:10:58 -0800 Received: from snort.melbourne.sgi.com (snort.melbourne.sgi.com [134.14.55.149]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB36ApA09610 for ; Sun, 2 Dec 2001 22:10:51 -0800 Received: (from tes@localhost) by snort.melbourne.sgi.com (SGI-8.9.3/8.9.3) id RAA51470 for linux-xfs@oss.sgi.com; Mon, 3 Dec 2001 17:09:34 +1100 (EST) Date: Mon, 3 Dec 2001 17:09:34 +1100 (EST) From: Timothy Shimmin Message-Id: <200112030609.RAA51470@snort.melbourne.sgi.com> To: linux-xfs@oss.sgi.com Subject: TAKE - xfstests/064 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Ensure that dump quota checking is done for 064 (hardlink cumulative restore test). --Tim Date: Sun Dec 2 22:08:15 PST 2001 Workarea: snort.melbourne.sgi.com:/home/diskb/build4/tes/slinx-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107626a cmd/xfstests/common.dump - 1.24 cmd/xfstests/064 - 1.2 - Fix up 064 for quota checks. From owner-linux-xfs@oss.sgi.com Mon Dec 3 07:51:45 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB3Fpjt28492 for linux-xfs-outgoing; Mon, 3 Dec 2001 07:51:45 -0800 Received: from starship.berlin (dsl-213-023-038-044.arcor-ip.net [213.23.38.44]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB3Fpeo28470 for ; Mon, 3 Dec 2001 07:51:40 -0800 Received: from daniel by starship.berlin with local (Exim 3.32 #1 (Debian)) id 16AuSr-0000HG-00; Mon, 03 Dec 2001 15:53:01 +0100 Content-Type: text/plain; charset="iso-8859-1" From: Daniel Phillips To: Nathan Scott Subject: Re: [RFC][PATCH] VFS interface for extended attributes Date: Mon, 3 Dec 2001 15:52:58 +0100 X-Mailer: KMail [version 1.3.2] Cc: Alexander Viro , Andi Kleen , Andreas Gruenbacher , Linus Torvalds , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com References: <20011203115400.F39338@wobbly.melbourne.sgi.com> In-Reply-To: <20011203115400.F39338@wobbly.melbourne.sgi.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On December 3, 2001 01:54 am, Nathan Scott wrote: > ...BTW, we have reworked the interfaces once more and > will send out the latest revision in the next couple of days - > it does away with commands and flags completely, except for this > one instance of flags in the set operation... OK, well I can see some patterns emerging already: long sys_getxattr(char *path, char *name, void *value, size_t size, int flags) long sys_setxattr(char *path, char *name, void *value, size_t size, int flags) long sys_listxattr(char *path, char *name, void *value, size_t size, int flags) long sys_fgetxattr(int fd, char *name, void *value, size_t size, int flags) long sys_fsetxattr(int fd, char *name, void *value, size_t size, int flags) long sys_flistxattr(int fd, char *name, void *value, size_t size, int flags) Why don't I see 'delxattr'? Why is there a need for separate 'path' and 'fd' variants? Is there any other kind of 'attr' in the syscall interface? Why not spell it 'attr' instead of 'xaddr'? How about geta, seta, dela, lista? The idea of attribute class (namespace) isn't explicitly accomodated. I presume the intention is to encode the class as part of the attribute name and have the filesystem or vfs parse it out. Is that such a good idea? Why not pass the class explicitly and worry about the namespace parsing in user space? As far as listing attributes goes, is there ever a reason to list system and user attributes at the same time? IOW, the listxattr call needs a class parameter too. It doesn't name a 'name', at least if you accept my argument that the class should not be parsed inside the kernel. There's no particular reason to force all the parameter lists to be the same is there? -- Daniel From owner-linux-xfs@oss.sgi.com Mon Dec 3 09:58:21 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB3HwL505728 for linux-xfs-outgoing; Mon, 3 Dec 2001 09:58:21 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB3HwAo05704 for ; Mon, 3 Dec 2001 09:58:12 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id RAA1221330 for ; Mon, 3 Dec 2001 17:58:06 +0100 (CET) mail_from (sandeen@sgi.com) Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id KAA3739567; Mon, 3 Dec 2001 10:56:50 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id KAA13145; Mon, 3 Dec 2001 10:56:49 -0600 (CST) Subject: inter-mezzo From: Eric Sandeen To: oluap@definitylinux.com.br Cc: linux-xfs@oss.sgi.com In-Reply-To: <200112031450.fB3EoAm21222@oss.sgi.com> References: <200112031450.fB3EoAm21222@oss.sgi.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 03 Dec 2001 10:50:39 -0600 Message-Id: <1007398240.1944.8.camel@stout.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hm, I thought I had that sorted out. Let me take a look at it today, I'll get things fixed up. BTW, inter-mezzo is not currently working with XFS, although we'll be able to get it to co-exist in the kernel. -Eric > When I'm trying to compile the kernel (2.4.16) from cvs it stops whith erros and reports this message: > > vfs.c: In function `lento_setattr': > vfs.c:321: parse error before `posix_acl_t' > vfs.c:321: warning: function declaration isn't a prototype > vfs.c:374: structure has no member named `set_posix_acl' > vfs.c:375: structure has no member named `set_posix_acl' > vfs.c:384: structure has no member named `set_posix_acl' > make[2]: ** [vfs.o] Erro 1 > make[1]: ** [_modsubdir_intermezzo] Erro 2 > make: ** [_mod_fs] Erro 2 > > Any body can help me ? -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Mon Dec 3 11:01:02 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB3J12408408 for linux-xfs-outgoing; Mon, 3 Dec 2001 11:01:02 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB3J0ro08379 for ; Mon, 3 Dec 2001 11:00:53 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id KAA09251 for ; Mon, 3 Dec 2001 10:00:51 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3742429 for ; Mon, 3 Dec 2001 11:59:35 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id LAA96789 for ; Mon, 3 Dec 2001 11:59:35 -0600 (CST) From: Steve Lord Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB3HpPR28246; Mon, 3 Dec 2001 11:51:25 -0600 Message-Id: <200112031751.fB3HpPR28246@jen.americas.sgi.com> Date: Mon, 3 Dec 2001 11:51:25 -0600 Subject: TAKE - restrict xfs inodes to 32 bits Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk It has always been possible for inode numbers in XFS to be more than 32 bits in size. This creates large problems for the rest of linux. This change restricts the allocation of inodes in an xfs filesystem to that part of the fs which will keep the inodes within 32 bits (basically the bottom 1 Tbyte with default mkfs parameters). For these large filesystems it also makes xfs file data prefer being higher up the filesystem to retain space for inodes and metadata in the lower portion. There is no on disk format change for this, but you will need to rebuild mkfs to build a filesystem which can use it - there is some automatic inode sizing code in mkfs which worked around the problem. Date: Mon Dec 3 09:52:09 PST 2001 Workarea: jen.americas.sgi.com:/src/lord/xfs-linux.2.4 The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107651a linux/fs/xfs/xfsidbg.c - 1.166 - Add new fields to perag dump linux/fs/xfs/xfs_ialloc.c - 1.150 - For inode allocation skip allocation groups which would place an inode out of range and make it larger than 32 bits. linux/fs/xfs/xfs_ag.h - 1.40 - Add new fields to perag structure linux/fs/xfs/xfs_vfsops.c - 1.328 - Add in XFS_MOUNT_32BITINODES mount option linux/fs/xfs/xfs_clnt.h - 1.25 - Define XFSMNT_32BITINODES flag linux/fs/xfs/xfs_mount.h - 1.131 - Prototype for xfs_initialize_perag linux/fs/xfs/xfs_mount.c - 1.264 - Add xfs_initialize_perag to calculate the maximum ag which is safe for inodes when operating in 32 bit mode. linux/fs/xfs/xfs_alloc.c - 1.144 - Changes for rotating file data through allocation groups when 32 bit inode code is enabled. linux/fs/xfs/xfs_alloc.h - 1.48 - New flags for data allocation mode linux/fs/xfs/xfs_fsops.c - 1.71 - Reinitialize the xfs_initialize_perag after a growfs linux/fs/xfs/xfs_bmap.c - 1.278 - Pass in new flag to allocator indicating if this is the initial data allocation for a file or not. linux/fs/xfs/linux/xfs_super.c - 1.145 - Add 32 bit inode flag to mount options cmd/xfsprogs/mkfs/xfs_mkfs.c - 1.17 - Comment out automatic inode size selection code in mkfs From owner-linux-xfs@oss.sgi.com Mon Dec 3 12:32:08 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB3KW8Q13505 for linux-xfs-outgoing; Mon, 3 Dec 2001 12:32:08 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB3KW1o13481 for ; Mon, 3 Dec 2001 12:32:02 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id UAA1308260 for ; Mon, 3 Dec 2001 20:31:55 +0100 (CET) mail_from (eric@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id NAA3756817 for ; Mon, 3 Dec 2001 13:30:32 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id NAA12596 for ; Mon, 3 Dec 2001 13:30:32 -0600 (CST) Received: by stout.americas.sgi.com (8.11.6/SGI-client-1.7) id fB3JOL310088; Mon, 3 Dec 2001 13:24:21 -0600 Message-Id: <200112031924.fB3JOL310088@stout.americas.sgi.com> Date: Mon, 3 Dec 2001 13:24:21 -0600 From: Eric Sandeen Subject: TAKE - work around acl/intermezzo problems. Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk This is more of a work-around than a fix, intermezzo looks for CONFIG_FS_POSIX_ACL in the config, which is part of the XFS patch, but intermezzo is really looking for the bestbits.at implementation, so things won't compile. Since you can't apply the bestbits code to the XFS tree at the moment anyway, this just turns off acl support in intermezzo to avoid the problem. Some day this will all come together in a grand unified acl implementation. :) Date: Mon Dec 3 11:27:52 PST 2001 Workarea: stout.americas.sgi.com:/localhome/eric/2.4.x-xfs/workarea-reallyclean The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107666a linux/fs/intermezzo/vfs.c - 1.3 - Work around ACL implementation clash From owner-linux-xfs@oss.sgi.com Mon Dec 3 12:58:07 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB3Kw7r14569 for linux-xfs-outgoing; Mon, 3 Dec 2001 12:58:07 -0800 Received: from mailhost.idcomm.com (mailhost.idcomm.com [207.40.196.14]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB3Kvwo14546 for ; Mon, 3 Dec 2001 12:57:58 -0800 Received: from idcomm.com (IDENT:PLRZg3qmWJqogTtgb/7/MdxUUUITWvjm@k56-pip39.idcomm.com [209.60.72.166]) by mailhost.idcomm.com (8.10.2/8.10.0) with ESMTP id fB3JxDE07241 for ; Mon, 3 Dec 2001 12:59:13 -0700 Message-ID: <3C0BD984.DF23699B@idcomm.com> Date: Mon, 03 Dec 2001 12:59:00 -0700 From: "D. Stimits" Reply-To: stimits@idcomm.com X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.6-pre1-xfs-4 i686) X-Accept-Language: en MIME-Version: 1.0 To: "XFS: linux-xfs@oss.sgi.com" Subject: Re: TAKE - restrict xfs inodes to 32 bits References: <200112031751.fB3HpPR28246@jen.americas.sgi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Steve Lord wrote: > > It has always been possible for inode numbers in XFS to be more than 32 > bits in size. This creates large problems for the rest of linux. This > change restricts the allocation of inodes in an xfs filesystem to that > part of the fs which will keep the inodes within 32 bits (basically > the bottom 1 Tbyte with default mkfs parameters). For these large > filesystems it also makes xfs file data prefer being higher up the > filesystem to retain space for inodes and metadata in the lower > portion. > > There is no on disk format change for this, but you will need to rebuild > mkfs to build a filesystem which can use it - there is some automatic > inode sizing code in mkfs which worked around the problem. Is there any chance a current filesystem on x86 2.4.x linux, disk sizes under 100 GB, would already have nodes greater than 32 bit? I'm assuming that upgrading won't break existing filesystems, i.e., do I need to fear using the 32 bit restrictions by mixing it with my current filesystem? I'm assuming that any ordinary size system (less than 1TB) will never have an inode above the 32 bit boundary. If not, I'd love a tool to sample all inodes and report the highest inode number, along with a simple "Exceeds 32 bits" or not type statement. D. Stimits, stimits@idcomm.com > > Date: Mon Dec 3 09:52:09 PST 2001 > Workarea: jen.americas.sgi.com:/src/lord/xfs-linux.2.4 > > The following file(s) were checked into: > bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs > > Modid: 2.4.x-xfs:slinx:107651a > linux/fs/xfs/xfsidbg.c - 1.166 > - Add new fields to perag dump > > linux/fs/xfs/xfs_ialloc.c - 1.150 > - For inode allocation skip allocation groups which would place an inode > out of range and make it larger than 32 bits. > > linux/fs/xfs/xfs_ag.h - 1.40 > - Add new fields to perag structure > > linux/fs/xfs/xfs_vfsops.c - 1.328 > - Add in XFS_MOUNT_32BITINODES mount option > > linux/fs/xfs/xfs_clnt.h - 1.25 > - Define XFSMNT_32BITINODES flag > > linux/fs/xfs/xfs_mount.h - 1.131 > - Prototype for xfs_initialize_perag > > linux/fs/xfs/xfs_mount.c - 1.264 > - Add xfs_initialize_perag to calculate the maximum ag which is safe for > inodes when operating in 32 bit mode. > > linux/fs/xfs/xfs_alloc.c - 1.144 > - Changes for rotating file data through allocation groups when > 32 bit inode code is enabled. > > linux/fs/xfs/xfs_alloc.h - 1.48 > - New flags for data allocation mode > > linux/fs/xfs/xfs_fsops.c - 1.71 > - Reinitialize the xfs_initialize_perag after a growfs > > linux/fs/xfs/xfs_bmap.c - 1.278 > - Pass in new flag to allocator indicating if this is the initial > data allocation for a file or not. > > linux/fs/xfs/linux/xfs_super.c - 1.145 > - Add 32 bit inode flag to mount options > > cmd/xfsprogs/mkfs/xfs_mkfs.c - 1.17 > - Comment out automatic inode size selection code in mkfs From owner-linux-xfs@oss.sgi.com Mon Dec 3 13:08:14 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB3L8Eu16278 for linux-xfs-outgoing; Mon, 3 Dec 2001 13:08:14 -0800 Received: from rj.sgi.com (rj.SGI.COM [204.94.215.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB3L88o16254 for ; Mon, 3 Dec 2001 13:08:08 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by rj.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB3K82Y22821 for ; Mon, 3 Dec 2001 12:08:02 -0800 Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id OAA3752423; Mon, 3 Dec 2001 14:06:46 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id OAA02911; Mon, 3 Dec 2001 14:06:46 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB3K6hO28520; Mon, 3 Dec 2001 14:06:43 -0600 Subject: Re: TAKE - restrict xfs inodes to 32 bits From: Steve Lord To: stimits@idcomm.com Cc: "XFS: ""linux-xfs@oss.sgi.com" In-Reply-To: <3C0BD984.DF23699B@idcomm.com> References: <200112031751.fB3HpPR28246@jen.americas.sgi.com> <3C0BD984.DF23699B@idcomm.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 03 Dec 2001 14:06:43 -0600 Message-Id: <1007410003.28378.0.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Mon, 2001-12-03 at 13:59, D. Stimits wrote: > Steve Lord wrote: > > > > It has always been possible for inode numbers in XFS to be more than 32 > > bits in size. This creates large problems for the rest of linux. This > > change restricts the allocation of inodes in an xfs filesystem to that > > part of the fs which will keep the inodes within 32 bits (basically > > the bottom 1 Tbyte with default mkfs parameters). For these large > > filesystems it also makes xfs file data prefer being higher up the > > filesystem to retain space for inodes and metadata in the lower > > portion. > > > > There is no on disk format change for this, but you will need to rebuild > > mkfs to build a filesystem which can use it - there is some automatic > > inode sizing code in mkfs which worked around the problem. > > Is there any chance a current filesystem on x86 2.4.x linux, disk sizes > under 100 GB, would already have nodes greater than 32 bit? I'm assuming > that upgrading won't break existing filesystems, i.e., do I need to fear > using the 32 bit restrictions by mixing it with my current filesystem? > I'm assuming that any ordinary size system (less than 1TB) will never > have an inode above the 32 bit boundary. If not, I'd love a tool to > sample all inodes and report the highest inode number, along with a > simple "Exceeds 32 bits" or not type statement. > > D. Stimits, stimits@idcomm.com > No, there is no chance - it takes a 1Tbyte filesystem for the overflow to occur. Steve -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Mon Dec 3 16:16:20 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB40GK327081 for linux-xfs-outgoing; Mon, 3 Dec 2001 16:16:20 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB40G8o27047 for ; Mon, 3 Dec 2001 16:16:08 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via SMTP id PAA01840 for ; Mon, 3 Dec 2001 15:15:54 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id KAA14336; Tue, 4 Dec 2001 10:14:43 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id KAA44348; Tue, 4 Dec 2001 10:14:40 +1100 (AEDT) Date: Tue, 4 Dec 2001 10:14:39 +1100 From: Nathan Scott To: Daniel Phillips Cc: Alexander Viro , Andi Kleen , Andreas Gruenbacher , Linus Torvalds , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [RFC][PATCH] VFS interface for extended attributes Message-ID: <20011204101439.E43856@wobbly.melbourne.sgi.com> References: <20011203115400.F39338@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from phillips@bonn-fries.net on Mon, Dec 03, 2001 at 03:52:58PM +0100 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi Daniel, On Mon, Dec 03, 2001 at 03:52:58PM +0100, Daniel Phillips wrote: > On December 3, 2001 01:54 am, Nathan Scott wrote: > > ...BTW, we have reworked the interfaces once more and > > will send out the latest revision in the next couple of days - > > it does away with commands and flags completely, except for this > > one instance of flags in the set operation... > > OK, well I can see some patterns emerging already: > > long sys_getxattr(char *path, char *name, void *value, size_t size, int flags) > long sys_setxattr(char *path, char *name, void *value, size_t size, int flags) > long sys_listxattr(char *path, char *name, void *value, size_t size, int flags) Not sure where you got those from (one of the early patches?), but it looks more like this now: long sys_setxattr(char *path, char *name, void *value, size_t size, int flags); long sys_getxattr(char *path, char *name, void *value, size_t size); long sys_listxattr(char *path, char *list, size_t size); > Why don't I see 'delxattr'? > Hmm.. good question. I had been achieving this through a flag to `set', but was never real happy about that - I guess a separate remove operation is the cleaner interface. > Why is there a need for separate 'path' and 'fd' variants? Applications - the ready example is the POSIX ACL API, which specifies path and file descriptor variants of its functions. > Is there any other kind of 'attr' in the syscall interface? > Why not spell it 'attr' instead of 'xaddr'? How about geta, seta, > dela, lista? We have "attr" inode_operations in the VFS, so the idea is to add... int (*revalidate) (struct dentry *); int (*setattr) (struct dentry *, struct iattr *); int (*getattr) (struct dentry *, struct iattr *); + int (*setxattr) (struct dentry *, char *, void *, size_t, int); + int (*getxattr) (struct dentry *, char *, void *, size_t); + int (*listxattr) (struct dentry *, char *, size_t); }; And I think I'll now separate out the remove operation too - something like: + int (*removexattr) (struct dentry *, char *); The system call names have been kept consistent with this VFS naming convention. > The idea of attribute class (namespace) isn't explicitly accomodated. > I presume the intention is to encode the class as part of the > attribute name and have the filesystem or vfs parse it out. That's correct. > Is that such a good idea? It simplifies things. It has worked for Andreas for awhile (its an idea directly from his current implementation), and I don't have any problems with it - it has proven quite simple to implement in XFS. > Why not pass the > class explicitly and worry about the namespace parsing in user space? > > As far as listing attributes goes, is there ever a reason to list system > and user attributes at the same time? For this area of the design, we stuck with Andreas' implementation. I suspect one of the main reasons Andreas did it this way is to be able to get all attribute names at once - for efficiency in backup programs, simplicity in file manager type programs, etc. > IOW, the listxattr call needs a class > parameter too. It doesn't name a 'name', at least if you accept my > argument that the class should not be parsed inside the kernel. Not sure I do - we had an initial version which separated name and namespace, but we eventually scrapped it. In particular, we ended up wanting this all-at-once list operation, which becomes much simpler when using a "fully-qualified" name approach. > There's no particular > reason to force all the parameter lists to be the same is there? No, I think you're looking at an early, draft version of the API. See the XFS CVS tree - cmd/attr2/man/* in particular. I'll post new patches here soon, just waiting to hear back from Andreas on a couple of issues (he's been away from mail for a few days). cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Mon Dec 3 17:30:47 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB41UlL32087 for linux-xfs-outgoing; Mon, 3 Dec 2001 17:30:47 -0800 Received: from zok.sgi.com (zok.sgi.com [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB41Uio32064 for ; Mon, 3 Dec 2001 17:30:44 -0800 Received: from snort.melbourne.sgi.com (snort.melbourne.sgi.com [134.14.55.149]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB40UbA08267 for ; Mon, 3 Dec 2001 16:30:37 -0800 Received: (from tes@localhost) by snort.melbourne.sgi.com (SGI-8.9.3/8.9.3) id LAA81632 for linux-xfs@oss.sgi.com; Tue, 4 Dec 2001 11:29:20 +1100 (EST) Date: Tue, 4 Dec 2001 11:29:20 +1100 (EST) From: Timothy Shimmin Message-Id: <200112040029.LAA81632@snort.melbourne.sgi.com> To: linux-xfs@oss.sgi.com Subject: TAKE - xfstests/common.dump Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Steve, please let me know if this fixes your dump qa failures (if not then I may need access to your machine). Thanks muchly, Tim. Date: Mon Dec 3 16:27:12 PST 2001 Workarea: snort.melbourne.sgi.com:/home/diskb/build4/tes/slinx-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107709a cmd/xfstests/common.dump - 1.25 - Ensure the awk passed in variables are quoted as they can contain spaces. Somehow this was not a problem on local testing machines. From owner-linux-xfs@oss.sgi.com Mon Dec 3 17:42:25 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB41gPK02363 for linux-xfs-outgoing; Mon, 3 Dec 2001 17:42:25 -0800 Received: from usazdolexch0.developonline.home (outlook.developonline.com [206.80.205.3]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB41gHo02338 for ; Mon, 3 Dec 2001 17:42:20 -0800 Received: from shiva.soulmachine.com ([172.21.1.96]) by usazdolexch0.developonline.home with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13) id XHDVF23R; Mon, 3 Dec 2001 17:42:08 -0700 Subject: Stable XFS+Kernel combo From: Blake Barnett To: linux-xfs@oss.sgi.com Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 03 Dec 2001 17:42:04 -0700 Message-Id: <1007426524.28458.8.camel@shiva> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk In the opinion of the XFS masters, what would be the most stable XFS-enabled kernel at this point? I've been using 2.4.9+XFS for a while and it seems pretty solid. But I'm wondering if one of the newer ones is working out very well. I'd like to get the benefits of the new VM, etc. I'm about to put a server in production with some large filesystems and I'd like to use XFS, I don't want to have to upgrade the kernel for a while on this box though... Suggestions/Opinions please! -- Blake Barnett (bdb) Sr. Unix Administrator DevelopOnline.com office: 480-377-6816 "Do, or do not. There is no try." --Yoda From owner-linux-xfs@oss.sgi.com Mon Dec 3 19:05:10 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB435AS08323 for linux-xfs-outgoing; Mon, 3 Dec 2001 19:05:10 -0800 Received: from zok.sgi.com (zok.sgi.com [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4356o08300 for ; Mon, 3 Dec 2001 19:05:06 -0800 Received: from relay1.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB4250A15460 for ; Mon, 3 Dec 2001 18:05:00 -0800 Received: from chuckle.americas.sgi.com (sandeen@chuckle.americas.sgi.com [128.162.211.44]) by relay1.corp.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id SAA05100; Mon, 3 Dec 2001 18:04:28 -0800 (PST) Date: Mon, 3 Dec 2001 20:04:28 -0600 (CST) From: Eric Sandeen X-X-Sender: To: Blake Barnett cc: Subject: Re: Stable XFS+Kernel combo In-Reply-To: <1007426524.28458.8.camel@shiva> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Well, we released 2.4.14 as "blessed" for XFS 1.0.2, that might be a place to start? -Eric On 3 Dec 2001, Blake Barnett wrote: > In the opinion of the XFS masters, what would be the most stable > XFS-enabled kernel at this point? I've been using 2.4.9+XFS for a > while and it seems pretty solid. But I'm wondering if one of the newer > ones is working out very well. I'd like to get the benefits of the new > VM, etc. > > I'm about to put a server in production with some large filesystems and > I'd like to use XFS, I don't want to have to upgrade the kernel for a > while on this box though... > > Suggestions/Opinions please! > > From owner-linux-xfs@oss.sgi.com Mon Dec 3 20:18:29 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB44ITt10334 for linux-xfs-outgoing; Mon, 3 Dec 2001 20:18:29 -0800 Received: from burgers.bubbanfriends.org (IDENT:postfix@burgers.bubbanfriends.org [216.140.122.113]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB44IPo10311 for ; Mon, 3 Dec 2001 20:18:25 -0800 Received: by burgers.bubbanfriends.org (Postfix, from userid 500) id 07758401A51; Mon, 3 Dec 2001 22:18:32 -0500 (EST) Received: from localhost (localhost [127.0.0.1]) by burgers.bubbanfriends.org (Postfix) with ESMTP id 01E64240022A; Mon, 3 Dec 2001 22:18:32 -0500 (EST) Date: Mon, 3 Dec 2001 22:18:31 -0500 (EST) From: Mike Burger To: Blake Barnett Cc: Subject: Re: Stable XFS+Kernel combo In-Reply-To: <1007426524.28458.8.camel@shiva> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk I'm using the 2.4.14 rpm from the oss.sgi site, and it's been working fine, so far. On 3 Dec 2001, Blake Barnett wrote: > In the opinion of the XFS masters, what would be the most stable > XFS-enabled kernel at this point? I've been using 2.4.9+XFS for a > while and it seems pretty solid. But I'm wondering if one of the newer > ones is working out very well. I'd like to get the benefits of the new > VM, etc. > > I'm about to put a server in production with some large filesystems and > I'd like to use XFS, I don't want to have to upgrade the kernel for a > while on this box though... > > Suggestions/Opinions please! > > From owner-linux-xfs@oss.sgi.com Mon Dec 3 22:08:41 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB468fV13138 for linux-xfs-outgoing; Mon, 3 Dec 2001 22:08:41 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB468do13113 for ; Mon, 3 Dec 2001 22:08:39 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via SMTP id VAA10992 for ; Mon, 3 Dec 2001 21:08:24 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id QAA16064 for ; Tue, 4 Dec 2001 16:07:18 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id QAA45151 for linux-xfs@oss.sgi.com; Tue, 4 Dec 2001 16:07:16 +1100 (AEDT) Date: Tue, 4 Dec 2001 16:07:16 +1100 From: Nathan Scott To: linux-xfs@oss.sgi.com Subject: Re: TAKE - work around acl/intermezzo problems. Message-ID: <20011204160715.H43856@wobbly.melbourne.sgi.com> References: <200112031924.fB3JOL310088@stout.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <200112031924.fB3JOL310088@stout.americas.sgi.com>; from sandeen@sgi.com on Mon, Dec 03, 2001 at 01:24:21PM -0600 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Mon, Dec 03, 2001 at 01:24:21PM -0600, Eric Sandeen wrote: > ... > Since you can't apply the bestbits code to the XFS tree at the moment > anyway, this just turns off acl support in intermezzo to avoid the > problem. Some day this will all come together in a grand unified > acl implementation. :) ... and there will be much rejoicing. -- Nathan From owner-linux-xfs@oss.sgi.com Mon Dec 3 23:35:10 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB47ZAT15477 for linux-xfs-outgoing; Mon, 3 Dec 2001 23:35:10 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB47Z7o15454 for ; Mon, 3 Dec 2001 23:35:07 -0800 Received: from sherman.melbourne.sgi.com (sherman.melbourne.sgi.com [134.14.55.175]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id WAA01285 for ; Mon, 3 Dec 2001 22:35:05 -0800 (PST) mail_from (kaos@sherman.melbourne.sgi.com) Received: (from kaos@localhost) by sherman.melbourne.sgi.com (8.9.3/8.9.3) id RAA07518; Tue, 4 Dec 2001 17:34:00 +1100 Date: Tue, 4 Dec 2001 17:34:00 +1100 From: Keith Owens Message-Id: <200112040634.RAA07518@sherman.melbourne.sgi.com> Subject: TAKE - Sync with kdb v1.9-2.4.16-i386 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Mon Dec 3 22:32:43 PST 2001 Workarea: sherman.melbourne.sgi.com:/build/kaos/2.4.x-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107720a linux/include/asm-um/kdb.h - 1.1 linux/Documentation/Configure.help - 1.120 linux/kdb/modules/kdbm_vm.c - 1.15 linux/include/linux/kdbprivate.h - 1.15 linux/include/linux/dis-asm.h - 1.10 linux/kdb/kdb_io.c - 1.11 linux/kdb/kdb_id.c - 1.11 linux/arch/i386/kdb/kdba_bt.c - 1.11 linux/kdb/ChangeLog - 1.13 From owner-linux-xfs@oss.sgi.com Mon Dec 3 23:51:19 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB47pJ215939 for linux-xfs-outgoing; Mon, 3 Dec 2001 23:51:19 -0800 Received: from shakti.rupa.com (foobar@cx838204-a.alsv1.occa.home.com [24.16.83.66]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB47pGo15916 for ; Mon, 3 Dec 2001 23:51:16 -0800 Received: by shakti.rupa.com (Postfix, from userid 500) id 17B8648FFC; Mon, 3 Dec 2001 22:51:13 -0800 (PST) To: linux-xfs@oss.sgi.com Subject: Re: TAKE - restrict xfs inodes to 32 bits References: <200112031751.fB3HpPR28246@jen.americas.sgi.com> <3C0BD984.DF23699B@idcomm.com> <1007410003.28378.0.camel@jen.americas.sgi.com> From: Rupa Schomaker Mail-Copies-To: never Date: Mon, 03 Dec 2001 22:51:12 -0800 In-Reply-To: <1007410003.28378.0.camel@jen.americas.sgi.com> (Steve Lord's message of "03 Dec 2001 14:06:43 -0600") Message-ID: Lines: 11 User-Agent: Gnus/5.090004 (Oort Gnus v0.04) XEmacs/21.4 (Copyleft, i386-debian-linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Steve Lord writes: > No, there is no chance - it takes a 1Tbyte filesystem for the overflow > to occur. Steve, What happens when one uses xfs_growfs to grow an existing < 1T filesystem to > 1T? -- -rupa From owner-linux-xfs@oss.sgi.com Tue Dec 4 07:06:19 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4F6JX02206 for linux-xfs-outgoing; Tue, 4 Dec 2001 07:06:19 -0800 Received: from mustard.heime.net (mustard.heime.net [194.234.65.222]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4F6Eo02180 for ; Tue, 4 Dec 2001 07:06:14 -0800 Received: from localhost (roy@localhost) by mustard.heime.net (8.11.6/8.9.3) with ESMTP id fB4E5tY20725 for ; Tue, 4 Dec 2001 15:05:55 +0100 Date: Tue, 4 Dec 2001 15:05:55 +0100 (CET) From: Roy Sigurd Karlsbakk X-Sender: To: XFS Mailing list Subject: using XFS for sequencial read from multiple large files Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi I want to sustain pretty high speed while reading files 1-5 GB in size. I also want to add / remove / replace these files every now and then. - How should I create the filesystem? Should I use the defaults? - How can I avoid fragmentation? The filesystem will probably be close-to full most of the time. I can probably live with quite some fragmentation, as long as the fragments are of a fixed size (say - 4MB). That way, I can increase the /proc/sys/vm/(min|max)-readahead to that value, reducing the effect of fragmentation. I beleive this might be true... - When trying out ReiserFS, I got a tip to mount the fs with 'notail'. Is there something like this in XFS? Thanks a lot roy -- Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA Computers are like air conditioners. They stop working when you open Windows. From owner-linux-xfs@oss.sgi.com Tue Dec 4 07:45:27 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4FjRQ04395 for linux-xfs-outgoing; Tue, 4 Dec 2001 07:45:27 -0800 Received: from maynard.mail.mindspring.net (maynard.mail.mindspring.net [207.69.200.243]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4FjNo04372 for ; Tue, 4 Dec 2001 07:45:23 -0800 Received: from user-uini6tt.dsl.mindspring.com ([165.121.27.189] helo=waltsathlon.localhost.net) by maynard.mail.mindspring.net with esmtp (Exim 3.33 #1) id 16BGoz-0008MK-00 for linux-xfs@oss.sgi.com; Tue, 04 Dec 2001 09:45:21 -0500 Received: from mindspring.com (localhost.localdomain [127.0.0.1]) by waltsathlon.localhost.net (Postfix) with ESMTP id C0A5BD040E38 for ; Tue, 4 Dec 2001 06:35:52 -0800 (PST) Message-ID: <3C0CDF48.6020704@mindspring.com> Date: Tue, 04 Dec 2001 06:35:52 -0800 From: Walt H User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.6+) Gecko/20011202 X-Accept-Language: en-us MIME-Version: 1.0 To: linux-xfs@oss.sgi.com Subject: Is xfs_fsr safe? Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hello all, I've been using xfs for nearly one year on one of my linux boxes, and after some minor glitches early on, all is working well. Great work! My question concerns xfs_fsr. Is it safe to use? Will I see any benefit if I set it up to run as a cron job maybe twice a week? I read through the man page and it sounds like something I should be doing for regular maintenance, I'm just a bit timid about reorganizing a mounted filesystem. The particular system I'm talking of is an Athlon XP 2.4.16 kernel compiled with kgcc running 2 IBM GXP60 40GB striped. Thanks in advance. -Walt From owner-linux-xfs@oss.sgi.com Tue Dec 4 08:12:47 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4GClw05242 for linux-xfs-outgoing; Tue, 4 Dec 2001 08:12:47 -0800 Received: from etoile.infra.idealx.com (sete.idealx.com [213.41.87.82]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4GCfo05220 for ; Tue, 4 Dec 2001 08:12:42 -0800 Received: (qmail 16455 invoked by uid 500); 4 Dec 2001 15:12:25 -0000 From: "Jérôme Tournier" Date: Tue, 4 Dec 2001 16:12:25 +0100 To: linux-xfs@oss.sgi.com Subject: patch anaconda XFS error Message-ID: <20011204161225.A8486@etoile.magasin1.auchan.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit User-Agent: Mutt/1.3.23i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hello, i use the anaconda package with XFS patch. I have already wrote here because when generating a new distribution, the installation proccess stopped because it can't load the xfs modules (depedencies problem). I still not resolved this problem (altought i add xfs_support and pagebuf in mk-images.i386 script). When i don't have this problem, i can continue but what we know found strange is that when installing the system in graphical mode, everything ok, but when installing with a boot disk, i have this error message: Traceback (innermost last): File "/usr/bin/anaconda", line 620, in ? intf.run(id, dispatch, configFileData) File "/var/tmp/anaconda-7.2//usr/lib/anaconda/text.py", line 386, in run File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 1004, in __call__ File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 948, in editCb File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 657, in editPartitionRequest File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 486, in fsOptionsDialog File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 251, in makeFsList File "/usr/lib/python1.5/snack.py", line 100, in setCurrent KeyError: Local variables in innermost frame: self: item: Any idea ? Thanks -- Jérôme From owner-linux-xfs@oss.sgi.com Tue Dec 4 08:29:36 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4GTaA05840 for linux-xfs-outgoing; Tue, 4 Dec 2001 08:29:36 -0800 Received: from mail.mpcf.com ([209.215.137.161]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4GTWo05818 for ; Tue, 4 Dec 2001 08:29:33 -0800 Received: from mpcf.com (10.0.0.8 [10.0.0.8]) by mail.mpcf.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13) id XSCXYX7C; Tue, 4 Dec 2001 10:24:17 -0500 Message-ID: <3C0CEAA2.43EF18C7@mpcf.com> Date: Tue, 04 Dec 2001 10:24:18 -0500 From: Kevin X-Mailer: Mozilla 4.79 [en] (X11; U; Linux 2.4.16 i686) X-Accept-Language: en MIME-Version: 1.0 To: linux-xfs@oss.sgi.com Subject: is XFS (on linux) safe for a production file server? Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk I've already set up a TurboLinux 6.5/Samba 2.2.2 box on a IBM x342 running XFS on top of hardware RAID5. It has been up for about two weeks now and seems to be running fine, but i have had occasional file problems. Files coming up corrupt with no obvious explanations as to why. My question is this .... is linux based XFS considered a production level product or is it still "development code"? I created the file system while running XFS 1.0.1 but i just bumped my kernel (2.4.14) up to XFS 1.0.2 Just looking for reassurance that XFS is OK..... :) Thanks, VeKTeReX From owner-linux-xfs@oss.sgi.com Tue Dec 4 08:37:18 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4GbI606142 for linux-xfs-outgoing; Tue, 4 Dec 2001 08:37:18 -0800 Received: from chaos.egr.duke.edu (chaos.egr.duke.edu [152.3.195.82]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4GbEo06119 for ; Tue, 4 Dec 2001 08:37:14 -0800 Received: from localhost (jlb@localhost) by chaos.egr.duke.edu (8.11.6/8.11.6) with ESMTP id fB4FbCd12962; Tue, 4 Dec 2001 10:37:12 -0500 X-Authentication-Warning: chaos.egr.duke.edu: jlb owned process doing -bs Date: Tue, 4 Dec 2001 10:37:12 -0500 (EST) From: Joshua Baker-LePain X-X-Sender: To: Kevin cc: Subject: Re: is XFS (on linux) safe for a production file server? In-Reply-To: <3C0CEAA2.43EF18C7@mpcf.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, 4 Dec 2001 at 10:24am, Kevin wrote > I've already set up a TurboLinux 6.5/Samba 2.2.2 box on > a IBM x342 running XFS on top of hardware RAID5. It > has been up for about two weeks now and seems to be > running fine, but i have had occasional file problems. > Files coming up corrupt with no obvious explanations as > to why. My question is this .... is linux based XFS > considered a production level product or is it still > "development code"? It's most certainly production level code. I've got 560GB on hardware RAID 5 in production since April with no XFS problems. Many other sites have similar (and much larger) amounts of data on it. If you're seeing corruption, file a full bug report here. Of most interest is what compiler did you use? -- Joshua Baker-LePain Department of Biomedical Engineering Duke University From owner-linux-xfs@oss.sgi.com Tue Dec 4 09:39:25 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4HdPE28168 for linux-xfs-outgoing; Tue, 4 Dec 2001 09:39:25 -0800 Received: from mustard.heime.net (mustard.heime.net [194.234.65.222]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4HdLo28146 for ; Tue, 4 Dec 2001 09:39:21 -0800 Received: from localhost (roy@localhost) by mustard.heime.net (8.11.6/8.9.3) with ESMTP id fB4Gcub21434; Tue, 4 Dec 2001 17:38:56 +0100 Date: Tue, 4 Dec 2001 17:38:56 +0100 (CET) From: Roy Sigurd Karlsbakk X-Sender: To: Joshua Baker-LePain cc: Kevin , Subject: Re: is XFS (on linux) safe for a production file server? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk > It's most certainly production level code. I've got 560GB on hardware > RAID 5 in production since April with no XFS problems. Many other sites > have similar (and much larger) amounts of data on it. If you're seeing > corruption, file a full bug report here. Of most interest is what > compiler did you use? hi What compiler should I use? -- Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA Computers are like air conditioners. They stop working when you open Windows. From owner-linux-xfs@oss.sgi.com Tue Dec 4 09:41:45 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4Hfj130490 for linux-xfs-outgoing; Tue, 4 Dec 2001 09:41:45 -0800 Received: from chaos.egr.duke.edu (chaos.egr.duke.edu [152.3.195.82]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4Hfgo30028 for ; Tue, 4 Dec 2001 09:41:42 -0800 Received: from localhost (jlb@localhost) by chaos.egr.duke.edu (8.11.6/8.11.6) with ESMTP id fB4GfWr13329; Tue, 4 Dec 2001 11:41:32 -0500 X-Authentication-Warning: chaos.egr.duke.edu: jlb owned process doing -bs Date: Tue, 4 Dec 2001 11:41:32 -0500 (EST) From: Joshua Baker-LePain X-X-Sender: To: Roy Sigurd Karlsbakk cc: Kevin , Subject: Re: is XFS (on linux) safe for a production file server? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, 4 Dec 2001 at 5:38pm, Roy Sigurd Karlsbakk wrote > What compiler should I use? > As the FAQ states, egcs-2.91.66 (aka kgcc on recent RedHat systems) is the recommended compiler for production systems. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University From owner-linux-xfs@oss.sgi.com Tue Dec 4 09:45:58 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4HjwC31675 for linux-xfs-outgoing; Tue, 4 Dec 2001 09:45:58 -0800 Received: from mustard.heime.net (mustard.heime.net [194.234.65.222]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4Hjro31653 for ; Tue, 4 Dec 2001 09:45:53 -0800 Received: from localhost (roy@localhost) by mustard.heime.net (8.11.6/8.9.3) with ESMTP id fB4GjZV21481 for ; Tue, 4 Dec 2001 17:45:35 +0100 Date: Tue, 4 Dec 2001 17:45:35 +0100 (CET) From: Roy Sigurd Karlsbakk X-Sender: To: XFS Mailing list Subject: error making real-time filesystem + logging question Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi all I'm trying to make a real-time filesystem for testing purposes onto a single disk. (yes - I know it's silly, but it's only for testing). I'm trying to do this on one drive with two partitions, and it all messes up. see below for more info. roy -- from /proc/partitions major minor #blocks name 3 69 19277937 hdb5 3 70 730926 hdb6 -- # mkfs.xfs -f -r extsize=4m,rtdev=/dev/hdb5 /dev/hdb6 meta-data=/dev/hdb6 isize=256 agcount=8, agsize=22842 blks data = bsize=4096 blocks=182731, imaxpct=25 = sunit=0 swidth=0 blks, unwritten=0 = imaxbits=0 naming =version 2 bsize=4096 log =internal log bsize=4096 blocks=1200 realtime =/dev/hdb5 extsz=4194304 blocks=4819484, rtextents=4706 mkfs.xfs: lseek64 to 19740605952 failed: Invalid argument but... # mkfs.xfs -f -r extsize=4m,rtdev=/dev/hdb6 /dev/hdb5 works fine... ? que? -- Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA Computers are like air conditioners. They stop working when you open Windows. From owner-linux-xfs@oss.sgi.com Tue Dec 4 10:10:23 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4IAN232282 for linux-xfs-outgoing; Tue, 4 Dec 2001 10:10:23 -0800 Received: from ADSL-Bergs.RZ.RWTH-Aachen.DE (adsl-bergs.rz.RWTH-Aachen.DE [137.226.80.218]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4IAIo32237 for ; Tue, 4 Dec 2001 10:10:19 -0800 Received: from [192.168.1.2] (helo=ralf) by ADSL-Bergs.RZ.RWTH-Aachen.DE with smtp (Exim 3.12 #1) id 16BJ59-0001TN-00; Tue, 04 Dec 2001 18:10:11 +0100 From: "Ralf G. R. Bergs" To: "XFS Mailing list" Cc: "Roy Sigurd Karlsbakk" Date: Tue, 04 Dec 2001 18:10:10 +0100 Reply-To: "Ralf G. R. Bergs" X-Mailer: PMMail 2000 Professional (2.20.2370) For Windows 2000 (5.0.2195;2) In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Subject: Re: error making real-time filesystem + logging question Message-Id: Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, 4 Dec 2001 17:45:35 +0100 (CET), Roy Sigurd Karlsbakk wrote: >I'm trying to make a real-time filesystem This is from http://oss.sgi.com/projects/xfs/102_caveats.html > Realtime volumes not implemented > > Realtime volumes are not yet tested in XFS Linux and should not be used. I guess that pretty much says it all... :-) -- Verkaufe Original-BMW-Raeder: L I N U X .~. http://adsl-bergs.rz.rwth-aachen.de/~rabe The Choice /V\ of a GNU /( )\ Generation ^^-^^ From owner-linux-xfs@oss.sgi.com Tue Dec 4 10:10:23 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4IANC32283 for linux-xfs-outgoing; Tue, 4 Dec 2001 10:10:23 -0800 Received: from fep01-app.kolumbus.fi (fep01-0.kolumbus.fi [193.229.0.41]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4IAIo32236 for ; Tue, 4 Dec 2001 10:10:18 -0800 Received: from there ([62.248.175.83]) by fep01-app.kolumbus.fi (InterMail vM.5.01.03.08 201-253-122-118-108-20010628) with SMTP id <20011204171015.HTRE2704.fep01-app.kolumbus.fi@there>; Tue, 4 Dec 2001 19:10:15 +0200 Content-Type: text/plain; charset="iso-8859-1" From: Hristo Grigorov To: Walt H , linux-xfs@oss.sgi.com Subject: Re: Is xfs_fsr safe? Date: Tue, 4 Dec 2001 19:11:56 +0200 X-Mailer: KMail [version 1.3.1] References: <3C0CDF48.6020704@mindspring.com> In-Reply-To: <3C0CDF48.6020704@mindspring.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <20011204171015.HTRE2704.fep01-app.kolumbus.fi@there> Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tuesday 04 December 2001 16:35, Walt H wrote: > Hello all, > > I've been using xfs for nearly one year on one of my linux boxes, and > after some minor glitches early on, all is working well. Great work! My > question concerns xfs_fsr. Is it safe to use? Will I see any benefit if > I set it up to run as a cron job maybe twice a week? I read through the > man page and it sounds like something I should be doing for regular > maintenance, I'm just a bit timid about reorganizing a mounted > filesystem. The particular system I'm talking of is an Athlon XP 2.4.16 > kernel compiled with kgcc running 2 IBM GXP60 40GB striped. Thanks in > advance. > > -Walt I don't know from engineering point of view is it safe, most probably yes. But I have it in my crontab (runned once a day) and I didn't noticed any problems so far. -- Cheers, Hristo. From owner-linux-xfs@oss.sgi.com Tue Dec 4 10:14:36 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4IEaT32573 for linux-xfs-outgoing; Tue, 4 Dec 2001 10:14:36 -0800 Received: from mustard.heime.net (mustard.heime.net [194.234.65.222]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4IEUo32551 for ; Tue, 4 Dec 2001 10:14:31 -0800 Received: from localhost (roy@localhost) by mustard.heime.net (8.11.6/8.9.3) with ESMTP id fB4HE9P21706; Tue, 4 Dec 2001 18:14:09 +0100 Date: Tue, 4 Dec 2001 18:14:09 +0100 (CET) From: Roy Sigurd Karlsbakk X-Sender: To: "Ralf G. R. Bergs" cc: XFS Mailing list Subject: Re: error making real-time filesystem + logging question In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk ok.. bummer! I really need to increase some sort of chunk size dramatically. Do you know any other way of doing this? I also want the log somewhere else, like on a separate spindle. I may be able to use ReiserFS to do this, but I don't know yet. roy On Tue, 4 Dec 2001, Ralf G. R. Bergs wrote: > On Tue, 4 Dec 2001 17:45:35 +0100 (CET), Roy Sigurd Karlsbakk wrote: > > >I'm trying to make a real-time filesystem > > This is from http://oss.sgi.com/projects/xfs/102_caveats.html > > > Realtime volumes not implemented > > > > Realtime volumes are not yet tested in XFS Linux and should not be used. > > I guess that pretty much says it all... :-) > > > > -- > Verkaufe Original-BMW-Raeder: L I N U X .~. > http://adsl-bergs.rz.rwth-aachen.de/~rabe The Choice /V\ > of a GNU /( )\ > Generation ^^-^^ > > -- Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA Computers are like air conditioners. They stop working when you open Windows. From owner-linux-xfs@oss.sgi.com Tue Dec 4 10:20:27 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4IKRx00306 for linux-xfs-outgoing; Tue, 4 Dec 2001 10:20:27 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4IKKo32751 for ; Tue, 4 Dec 2001 10:20:20 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id JAA03693 for ; Tue, 4 Dec 2001 09:20:11 -0800 (PST) mail_from (sandeen@sgi.com) Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3749737; Tue, 4 Dec 2001 11:18:55 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id LAA62227; Tue, 4 Dec 2001 11:18:54 -0600 (CST) Subject: Re: is XFS (on linux) safe for a production file server? From: Eric Sandeen To: Kevin Cc: linux-xfs@oss.sgi.com In-Reply-To: <3C0CEAA2.43EF18C7@mpcf.com> References: <3C0CEAA2.43EF18C7@mpcf.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 04 Dec 2001 11:12:35 -0600 Message-Id: <1007485955.12169.9.camel@stout.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi Roy - Many people are using it in production, see http://oss.sgi.com/projects/xfs/xfs_users.html If you're seeing "corruption," please send more details so we can figure out what's going on. Thanks, -Eric On Tue, 2001-12-04 at 09:24, Kevin wrote: > I've already set up a TurboLinux 6.5/Samba 2.2.2 box on > a IBM x342 running XFS on top of hardware RAID5. It > has been up for about two weeks now and seems to be > running fine, but i have had occasional file problems. > Files coming up corrupt with no obvious explanations as > to why. My question is this .... is linux based XFS > considered a production level product or is it still > "development code"? > > I created the file system while running XFS 1.0.1 but i > just bumped my kernel (2.4.14) up to XFS 1.0.2 > > Just looking for reassurance that XFS is OK..... :) > > Thanks, > VeKTeReX -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Tue Dec 4 10:21:01 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4IL1r00460 for linux-xfs-outgoing; Tue, 4 Dec 2001 10:21:01 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4IKso00434 for ; Tue, 4 Dec 2001 10:20:54 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id JAA03536 for ; Tue, 4 Dec 2001 09:20:48 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3757175; Tue, 4 Dec 2001 11:19:30 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id LAA71421; Tue, 4 Dec 2001 11:19:30 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB4HJJR16540; Tue, 4 Dec 2001 11:19:19 -0600 Subject: Re: error making real-time filesystem + logging question From: Steve Lord To: "Ralf G. R. Bergs" Cc: XFS Mailing list , Roy Sigurd Karlsbakk In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 04 Dec 2001 11:19:19 -0600 Message-Id: <1007486359.13856.5.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, 2001-12-04 at 11:10, Ralf G. R. Bergs wrote: > On Tue, 4 Dec 2001 17:45:35 +0100 (CET), Roy Sigurd Karlsbakk wrote: > > >I'm trying to make a real-time filesystem > > This is from http://oss.sgi.com/projects/xfs/102_caveats.html > > > Realtime volumes not implemented > > > > Realtime volumes are not yet tested in XFS Linux and should not be used. > > I guess that pretty much says it all... :-) Except that I am not convinced the caveats page is right in this respect. To make a realtime fs you will need two partitions, mkfs will need to be told about the data and realtime partitions, and mount will also need to be told about them. mkfs -t xfs -r rtdev=/dev/hdb1 /dev/hda1 mount -t xfs -o rtdev=/dev/hdb1 /dev/hda1 /xfs You also need special ioctl calls to mark a file as realtime before it will place data into the realtime subvolume: XFS_IOC_FSSETXATTR This takes struct fsxattr as the argument: struct fsxattr { __u32 fsx_xflags; /* xflags field value (get/set) */ __u32 fsx_extsize; /* extsize field value (get/set)*/ __u32 fsx_nextents; /* nextents field value (get) */ unsigned char fsx_pad[16]; }; You need to set the flag XFS_XFLAG_REALTIME in the fsx_xflags field, and specify the size of realtime extent you want by default in the extsize field. Steve p.s. I have not tried this recently - but it has worked in the past. > > > > -- > Verkaufe Original-BMW-Raeder: L I N U X .~. > http://adsl-bergs.rz.rwth-aachen.de/~rabe The Choice /V\ > of a GNU /( )\ > Generation ^^-^^ > -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Tue Dec 4 10:30:15 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4IUFt01077 for linux-xfs-outgoing; Tue, 4 Dec 2001 10:30:15 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4IUCo01054 for ; Tue, 4 Dec 2001 10:30:12 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id JAA23964 for ; Tue, 4 Dec 2001 09:29:58 -0800 (PST) mail_from (sandeen@sgi.com) Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3763869; Tue, 4 Dec 2001 11:28:53 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id LAA85719; Tue, 4 Dec 2001 11:28:52 -0600 (CST) Subject: Re: error making real-time filesystem + logging question From: Eric Sandeen To: Roy Sigurd Karlsbakk Cc: "Ralf G. R. Bergs" , XFS Mailing list In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 04 Dec 2001 11:22:33 -0600 Message-Id: <1007486554.12410.11.camel@stout.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, 2001-12-04 at 11:14, Roy Sigurd Karlsbakk wrote: > I also want the log somewhere else, like > on a separate spindle. Oh, and logging somewhere else is not a problem, just use the logdev argument to mkfs.xfs and mount: mkfs.xfs -l logdev=/dev/sdb1,size=10000b /dev/sda1 mount -o logdev=/dev/sdb1 /dev/sda1 /mnt/point -Eric -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Tue Dec 4 10:31:14 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4IVEq01230 for linux-xfs-outgoing; Tue, 4 Dec 2001 10:31:14 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4IV8o01201 for ; Tue, 4 Dec 2001 10:31:08 -0800 Received: from ledzep.americas.sgi.com (relay.sgi.com [137.38.226.97] (may be forged)) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id JAA08294 for ; Tue, 4 Dec 2001 09:31:06 -0800 (PST) mail_from (nstraz@sgi.com) Received: from maine.americas.sgi.com (maine.americas.sgi.com [128.162.191.42]) by ledzep.americas.sgi.com (SGI-SGI-8.9.3/americas-smart-nospam1.1) with ESMTP id LAA77611; Tue, 4 Dec 2001 11:27:12 -0600 (CST) Received: from nstraz by maine.americas.sgi.com with local (Exim 3.33 #1 (Debian)) id 16BJO9-0000EX-00; Tue, 04 Dec 2001 11:29:49 -0600 Date: Tue, 4 Dec 2001 11:29:49 -0600 From: Nathan Straz To: Roy Sigurd Karlsbakk Cc: "Ralf G. R. Bergs" , XFS Mailing list Subject: Re: error making real-time filesystem + logging question Message-ID: <20011204112949.U1046@sgi.com> Mail-Followup-To: Roy Sigurd Karlsbakk , "Ralf G. R. Bergs" , XFS Mailing list References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.3.23i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, Dec 04, 2001 at 06:14:09PM +0100, Roy Sigurd Karlsbakk wrote: > On Tue, 4 Dec 2001, Ralf G. R. Bergs wrote: > > > On Tue, 4 Dec 2001 17:45:35 +0100 (CET), Roy Sigurd Karlsbakk wrote: > > > > >I'm trying to make a real-time filesystem > > > > This is from http://oss.sgi.com/projects/xfs/102_caveats.html > > > > > Realtime volumes not implemented > > > > > > Realtime volumes are not yet tested in XFS Linux and should not be used. > > > > I guess that pretty much says it all... :-) > ok.. bummer! > > I really need to increase some sort of chunk size dramatically. Do you > know any other way of doing this? I also want the log somewhere else, like > on a separate spindle. I may be able to use ReiserFS to do this, but I > don't know yet. Realtime subvolumes is on my things to test before the next release. I don't know when it will get done, but it will take a bit of test development to do it. Realtime files on XFS require an ioctl, so apps will need some work to support them. To create an external log, you should use something like: mkfs.xfs -l logdev=/dev/sdb1,size=10000b /dev/sda1 Where /dev/sda1 is your XFS partition and /dev/sdb1 is your log partition on a separate spindle. Adjust size= to fit your application. As far as "chunk size" is concerned, you probably want to look at something like "-d sunit=,swidth=". Read more about these in the mkfs.xfs man page. -- Nate Straz nstraz@sgi.com sgi, inc http://www.sgi.com/ Linux Test Project http://ltp.sf.net/ From owner-linux-xfs@oss.sgi.com Tue Dec 4 10:31:36 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4IVaQ01374 for linux-xfs-outgoing; Tue, 4 Dec 2001 10:31:36 -0800 Received: from mustard.heime.net (mustard.heime.net [194.234.65.222]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4IVQo01323 for ; Tue, 4 Dec 2001 10:31:26 -0800 Received: from localhost (roy@localhost) by mustard.heime.net (8.11.6/8.9.3) with ESMTP id fB4HV2B21832; Tue, 4 Dec 2001 18:31:02 +0100 Date: Tue, 4 Dec 2001 18:31:02 +0100 (CET) From: Roy Sigurd Karlsbakk X-Sender: To: Steve Lord cc: "Ralf G. R. Bergs" , XFS Mailing list Subject: Re: error making real-time filesystem + logging question In-Reply-To: <1007486359.13856.5.camel@jen.americas.sgi.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk ok... What's the main difference between a realtime partition and a data partiton? Can a datapartition (or non-realtime partition) be given an extsize= parameter? That's the one I REALLY need... On 4 Dec 2001, Steve Lord wrote: > On Tue, 2001-12-04 at 11:10, Ralf G. R. Bergs wrote: > > On Tue, 4 Dec 2001 17:45:35 +0100 (CET), Roy Sigurd Karlsbakk wrote: > > > > >I'm trying to make a real-time filesystem > > > > This is from http://oss.sgi.com/projects/xfs/102_caveats.html > > > > > Realtime volumes not implemented > > > > > > Realtime volumes are not yet tested in XFS Linux and should not be used. > > > > I guess that pretty much says it all... :-) > > Except that I am not convinced the caveats page is right in this > respect. To make a realtime fs you will need two partitions, mkfs > will need to be told about the data and realtime partitions, and > mount will also need to be told about them. > > mkfs -t xfs -r rtdev=/dev/hdb1 /dev/hda1 > > mount -t xfs -o rtdev=/dev/hdb1 /dev/hda1 /xfs > > You also need special ioctl calls to mark a file as realtime before it > will place data into the realtime subvolume: > > XFS_IOC_FSSETXATTR > > This takes struct fsxattr as the argument: > struct fsxattr { > __u32 fsx_xflags; /* xflags field value (get/set) > */ > __u32 fsx_extsize; /* extsize field value > (get/set)*/ > __u32 fsx_nextents; /* nextents field value (get) > */ > unsigned char fsx_pad[16]; > }; > > You need to set the flag XFS_XFLAG_REALTIME in the fsx_xflags field, and > specify the size of realtime extent you want by default in the extsize > field. > > Steve > > p.s. I have not tried this recently - but it has worked in the past. > > > > > > > > > > -- > > Verkaufe Original-BMW-Raeder: L I N U X .~. > > http://adsl-bergs.rz.rwth-aachen.de/~rabe The Choice /V\ > > of a GNU /( )\ > > Generation ^^-^^ > > > -- > > Steve Lord voice: +1-651-683-3511 > Principal Engineer, Filesystem Software email: lord@sgi.com > -- Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA Computers are like air conditioners. They stop working when you open Windows. From owner-linux-xfs@oss.sgi.com Tue Dec 4 10:36:25 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4IaPq01589 for linux-xfs-outgoing; Tue, 4 Dec 2001 10:36:25 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4IaLo01567 for ; Tue, 4 Dec 2001 10:36:21 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id JAA09454 for ; Tue, 4 Dec 2001 09:36:20 -0800 (PST) mail_from (roehrich@sgi.com) Received: from thistle-e185.americas.sgi.com (thistle-e185.americas.sgi.com [128.162.185.204]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3693610 for ; Tue, 4 Dec 2001 11:35:02 -0600 (CST) Received: from clink.americas.sgi.com (clink-eth.americas.sgi.com [128.162.2.8]) by thistle-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id LAA08320 for ; Tue, 4 Dec 2001 11:35:02 -0600 (CST) Received: (from roehrich@localhost) by clink.americas.sgi.com (SGI-8.9.3/8.9.3/erikj-IRIX) id LAA23008 for linux-xfs@oss.sgi.com; Tue, 4 Dec 2001 11:35:02 -0600 (CST) Date: Tue, 4 Dec 2001 11:35:02 -0600 (CST) From: Dean Roehrich Message-Id: <200112041735.LAA23008@clink.americas.sgi.com> To: linux-xfs@oss.sgi.com Subject: TAKE - add VFS_DMAPI_FSYS_VECTOR for dmapi module Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Tue Dec 4 09:34:40 PST 2001 Workarea: clink-eth.americas.sgi.com:/data/clink/a67/roehrich/2.4.x-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107737a linux/fs/xfs/xfs_dmapi.h - 1.18 linux/fs/xfs/xfs_dmapi.c - 1.43 - Make xfs_dm_get_fsys_vector() non-static. linux/fs/xfs/xfs_vfsops.c - 1.329 - Add xfs_dm_get_fsys_vector to xfs_vfsops. linux/fs/xfs/xfsdmapistubs.c - 1.9 - Add stub for xfs_dm_get_fsys_vector. linux/fs/xfs/dmapi/dmapi_mountinfo.c - 1.8 - Use VFS_DMAPI_FSYS_VECTOR. linux/fs/xfs/linux/xfs_vfs.h - 1.7 - Add VFS_DMAPI_FSYS_VECTOR From owner-linux-xfs@oss.sgi.com Tue Dec 4 11:03:42 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4J3gs02791 for linux-xfs-outgoing; Tue, 4 Dec 2001 11:03:42 -0800 Received: from mailhost.idcomm.com (mailhost.idcomm.com [207.40.196.14]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4J3Zo02768 for ; Tue, 4 Dec 2001 11:03:35 -0800 Received: from idcomm.com (IDENT:oDbV3dZI8Woo9nlcw1f6F5mInsDGMt5N@k56-pip62.idcomm.com [209.60.72.189]) by mailhost.idcomm.com (8.10.2/8.10.0) with ESMTP id fB4I4uE12899 for ; Tue, 4 Dec 2001 11:04:56 -0700 Message-ID: <3C0D1038.1FA2EF43@idcomm.com> Date: Tue, 04 Dec 2001 11:04:40 -0700 From: "D. Stimits" Reply-To: stimits@idcomm.com X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.6-pre1-xfs-4 i686) X-Accept-Language: en MIME-Version: 1.0 CC: XFS Mailing list Subject: Re: error making real-time filesystem + logging question References: <20011204112949.U1046@sgi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Nathan Straz wrote: > > On Tue, Dec 04, 2001 at 06:14:09PM +0100, Roy Sigurd Karlsbakk wrote: > > On Tue, 4 Dec 2001, Ralf G. R. Bergs wrote: > > > > > On Tue, 4 Dec 2001 17:45:35 +0100 (CET), Roy Sigurd Karlsbakk wrote: > > > > > > >I'm trying to make a real-time filesystem > > > > > > This is from http://oss.sgi.com/projects/xfs/102_caveats.html > > > > > > > Realtime volumes not implemented > > > > > > > > Realtime volumes are not yet tested in XFS Linux and should not be used. > > > > > > I guess that pretty much says it all... :-) > > ok.. bummer! > > > > I really need to increase some sort of chunk size dramatically. Do you > > know any other way of doing this? I also want the log somewhere else, like > > on a separate spindle. I may be able to use ReiserFS to do this, but I > > don't know yet. > > Realtime subvolumes is on my things to test before the next release. I > don't know when it will get done, but it will take a bit of test > development to do it. Realtime files on XFS require an ioctl, so apps > will need some work to support them. Now I'm curious, what's the basic purpose and performance difference between a real-time filesystem and a "normal" filesystem? Is it some sort of caching behavior change more suitable for streaming media? Just curious. D. Stimits, stimits@idcomm.com > > To create an external log, you should use something like: > > mkfs.xfs -l logdev=/dev/sdb1,size=10000b /dev/sda1 > > Where /dev/sda1 is your XFS partition and /dev/sdb1 is your log > partition on a separate spindle. Adjust size= to fit your application. > > As far as "chunk size" is concerned, you probably want to look at > something like "-d sunit=,swidth=". Read more about these in the > mkfs.xfs man page. > > -- > Nate Straz nstraz@sgi.com > sgi, inc http://www.sgi.com/ > Linux Test Project http://ltp.sf.net/ From owner-linux-xfs@oss.sgi.com Tue Dec 4 11:58:09 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4Jw9A05005 for linux-xfs-outgoing; Tue, 4 Dec 2001 11:58:09 -0800 Received: from rj.sgi.com (rj.SGI.COM [204.94.215.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4Jw1o04980 for ; Tue, 4 Dec 2001 11:58:02 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by rj.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB4IvtY24148 for ; Tue, 4 Dec 2001 10:57:55 -0800 Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id MAA3735708; Tue, 4 Dec 2001 12:56:38 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id MAA72249; Tue, 4 Dec 2001 12:56:38 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB4IuQW17531; Tue, 4 Dec 2001 12:56:26 -0600 Subject: Re: TAKE - restrict xfs inodes to 32 bits From: Steve Lord To: Rupa Schomaker Cc: linux-xfs@oss.sgi.com In-Reply-To: References: <200112031751.fB3HpPR28246@jen.americas.sgi.com> <3C0BD984.DF23699B@idcomm.com> <1007410003.28378.0.camel@jen.americas.sgi.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 04 Dec 2001 12:56:26 -0600 Message-Id: <1007492186.15738.9.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, 2001-12-04 at 00:51, Rupa Schomaker wrote: > Steve Lord writes: > > No, there is no chance - it takes a 1Tbyte filesystem for the overflow > > to occur. > > Steve, > > What happens when one uses xfs_growfs to grow an existing < 1T > filesystem to > 1T? Well, if I fix one line of code it will work, although there are some pathological cases where you could grow a full filesystem from just less that 1 Tbyte to 2 Tbytes and end up with lots of room and no inode space, although in practice this should not happen. Steve > > -- > -rupa -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Tue Dec 4 12:32:26 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4KWQL05952 for linux-xfs-outgoing; Tue, 4 Dec 2001 12:32:26 -0800 Received: from smtpzilla2.xs4all.nl (smtpzilla2.xs4all.nl [194.109.127.138]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4KWLo05927 for ; Tue, 4 Dec 2001 12:32:21 -0800 Received: from auto-nb1.xs4all.nl (213-84-100-130.adsl.xs4all.nl [213.84.100.130]) by smtpzilla2.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB4JWB5a034869; Tue, 4 Dec 2001 20:32:12 +0100 (CET) Message-Id: <4.3.2.7.2.20011204202823.02c0e960@pop.xs4all.nl> X-Sender: knuffie@pop.xs4all.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Tue, 04 Dec 2001 20:29:06 +0100 To: Roy Sigurd Karlsbakk , "Ralf G. R. Bergs" From: Seth Mos Subject: Re: error making real-time filesystem + logging question Cc: XFS Mailing list In-Reply-To: References: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 18:14 4-12-2001 +0100, Roy Sigurd Karlsbakk wrote: >ok.. bummer! > >I really need to increase some sort of chunk size dramatically. Do you >know any other way of doing this? I also want the log somewhere else, like >on a separate spindle. I may be able to use ReiserFS to do this, but I >don't know yet. XFS supports external logs. See the manpage for mkfs.xfs -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Tue Dec 4 12:51:17 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4KpHF06594 for linux-xfs-outgoing; Tue, 4 Dec 2001 12:51:17 -0800 Received: from mustard.heime.net (mustard.heime.net [194.234.65.222]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4KpCo06561 for ; Tue, 4 Dec 2001 12:51:12 -0800 Received: from mail.karlsbakk.net (mustard.heime.net [194.234.65.222]) by mustard.heime.net (8.11.6/8.9.3) with SMTP id fB4Jomc22703; Tue, 4 Dec 2001 20:50:48 +0100 Message-Id: <200112041950.fB4Jomc22703@mustard.heime.net> Date: Tue, 04 Dec 2001 19:50:48 +0000 From: Roy Sigurd Karlsbakk Reply-To: roy@karlsbakk.net To: lord@sgi.com, roy@karlsbakk.net Cc: rabe@RWTH-Aachen.DE, linux-xfs@oss.sgi.com, stimits@idcomm.com X-Mailer: phpGroupWare (http://www.phpgroupware.org) Subject: Re: error making real-time filesystem + logging question MIME-version: 1.0 Content-type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-description: Mail message body Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >The XFS realtime subvolume is a separate dataspace in the filesystem, it >is only used for files marked as realtime after they are created. There >is nothing in the realtime subvolume except file data, where as in the >normal XFS data subvolume there is a mix of metadata and file data. >Space allocation in the realtime subvolume is managed by a different >allocator which gives out space in multiples of a pre-specified size, >the allocator uses a binary chop approach to space allocation which >is designed to avoid fragmentation at the expense of efficient use of >space - the allocator should also be faster. > >The end result is that streaming I/O should behave better on realtime >files and on regular files. ok. But XFS realtime subvolumes are, as I understand, not yet supported on Linux platform. Are there any plans of supporting it soon? How is a file placed or retrieved on the realtime subvolume? Is it flagged as it's been created, or does the realtime-part and the data-part come up as separate volumes with separate mount points? >One feature which can be used on normal files which might help with >placement is preallocation of space. Take a look at the use of the >XFS_IOC_RESVSP64 ioctl in cmd/xfsprogs/mkfile/xfs_mkfile.c Is this available through /proc ? roy -- -- Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA Computers are like air conditioners. They stop working when you open Windows. From owner-linux-xfs@oss.sgi.com Tue Dec 4 12:57:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4Kv3Z06804 for linux-xfs-outgoing; Tue, 4 Dec 2001 12:57:03 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4Kuvo06782 for ; Tue, 4 Dec 2001 12:56:57 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id LAA05548 for ; Tue, 4 Dec 2001 11:56:44 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id NAA3765248; Tue, 4 Dec 2001 13:55:39 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id NAA87387; Tue, 4 Dec 2001 13:55:39 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB4JtR120187; Tue, 4 Dec 2001 13:55:27 -0600 Subject: Re: error making real-time filesystem + logging question From: Steve Lord To: roy@karlsbakk.net Cc: rabe@RWTH-Aachen.DE, linux-xfs@oss.sgi.com, stimits@idcomm.com In-Reply-To: <200112041950.fB4Jomc22703@mustard.heime.net> References: <200112041950.fB4Jomc22703@mustard.heime.net> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 04 Dec 2001 13:55:27 -0600 Message-Id: <1007495727.19823.14.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, 2001-12-04 at 13:50, Roy Sigurd Karlsbakk wrote: > >The XFS realtime subvolume is a separate dataspace in the filesystem, it > >is only used for files marked as realtime after they are created. There > >is nothing in the realtime subvolume except file data, where as in the > >normal XFS data subvolume there is a mix of metadata and file data. > >Space allocation in the realtime subvolume is managed by a different > >allocator which gives out space in multiples of a pre-specified size, > >the allocator uses a binary chop approach to space allocation which > >is designed to avoid fragmentation at the expense of efficient use of > >space - the allocator should also be faster. > > > >The end result is that streaming I/O should behave better on realtime > >files and on regular files. > > ok. But XFS realtime subvolumes are, as I understand, not yet supported on Linux > platform. Are there any plans of supporting it soon? > > How is a file placed or retrieved on the realtime subvolume? Is it flagged as it's > been created, or does the realtime-part and the data-part come up as separate volumes > with separate mount points? See my previous email - the web page is wrong, and all the details are in the mail message. > > >One feature which can be used on normal files which might help with > >placement is preallocation of space. Take a look at the use of the > >XFS_IOC_RESVSP64 ioctl in cmd/xfsprogs/mkfile/xfs_mkfile.c > > Is this available through /proc ? ??? No it is an ioctl interface to an xfs file, /proc has nothing to do with it. Steve > > roy > > -- > -- > Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA > > Computers are like air conditioners. > They stop working when you open Windows. -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Tue Dec 4 13:04:43 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4L4hZ14378 for linux-xfs-outgoing; Tue, 4 Dec 2001 13:04:43 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4L4co14350 for ; Tue, 4 Dec 2001 13:04:38 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id MAA03682 for ; Tue, 4 Dec 2001 12:04:38 -0800 (PST) mail_from (roehrich@sgi.com) Received: from thistle-e185.americas.sgi.com (thistle-e185.americas.sgi.com [128.162.185.204]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id OAA3749736 for ; Tue, 4 Dec 2001 14:03:20 -0600 (CST) Received: from clink.americas.sgi.com (clink-eth.americas.sgi.com [128.162.2.8]) by thistle-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id OAA23599 for ; Tue, 4 Dec 2001 14:03:20 -0600 (CST) Received: (from roehrich@localhost) by clink.americas.sgi.com (SGI-8.9.3/8.9.3/erikj-IRIX) id OAA54282 for linux-xfs@oss.sgi.com; Tue, 4 Dec 2001 14:03:20 -0600 (CST) Date: Tue, 4 Dec 2001 14:03:20 -0600 (CST) From: Dean Roehrich Message-Id: <200112042003.OAA54282@clink.americas.sgi.com> To: linux-xfs@oss.sgi.com Subject: TAKE - switch CONFIG_XFS_DMAPI to CONFIG_HAVE_XFS_DMAPI, prep for module Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk If I build dmapi as a module then CONFIG_XFS_DMAPI won't be set--it'll be CONFIG_XFS_DMAPI_MODULE instead. So switch to using CONFIG_HAVE_XFS_DMAPI to see if DMAPI is being used, whether dmapi is a module or not. Date: Tue Dec 4 12:02:50 PST 2001 Workarea: clink-eth.americas.sgi.com:/data/clink/a67/roehrich/2.4.x-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107752a linux/fs/Config.in - 1.72 linux/fs/xfs/Makefile - 1.129 linux/fs/xfs/xfs_log_recover.c - 1.215 linux/fs/xfs/linux/xfs_lrw.c - 1.117 linux/fs/xfs/linux/Makefile - 1.45 linux/fs/xfs/linux/xfs_file.c - 1.54 linux/fs/xfs/linux/xfs_super.h - 1.12 linux/fs/xfs/linux/xfs_super.c - 1.146 linux/fs/xfs/Makefile.in - 1.8 linux/fs/xfs/linux/Makefile.in - 1.4 - No Message Supplied From owner-linux-xfs@oss.sgi.com Tue Dec 4 14:44:43 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4MihM25534 for linux-xfs-outgoing; Tue, 4 Dec 2001 14:44:43 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4MiXo25508 for ; Tue, 4 Dec 2001 14:44:33 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id NAA09505 for ; Tue, 4 Dec 2001 13:44:25 -0800 (PST) mail_from (roehrich@sgi.com) Received: from thistle-e185.americas.sgi.com (thistle-e185.americas.sgi.com [128.162.185.204]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id PAA3741458 for ; Tue, 4 Dec 2001 15:43:09 -0600 (CST) Received: from clink.americas.sgi.com (clink-eth.americas.sgi.com [128.162.2.8]) by thistle-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id PAA60262 for ; Tue, 4 Dec 2001 15:43:08 -0600 (CST) Received: (from roehrich@localhost) by clink.americas.sgi.com (SGI-8.9.3/8.9.3/erikj-IRIX) id PAA90028 for linux-xfs@oss.sgi.com; Tue, 4 Dec 2001 15:43:08 -0600 (CST) Date: Tue, 4 Dec 2001 15:43:08 -0600 (CST) From: Dean Roehrich Message-Id: <200112042143.PAA90028@clink.americas.sgi.com> To: linux-xfs@oss.sgi.com Subject: TAKE - move fs/xfs/dmapi to fs/xfs_dmapi Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Tue Dec 4 13:42:31 PST 2001 Workarea: clink-eth.americas.sgi.com:/data/clink/a67/roehrich/2.4.x-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107764a linux/fs/xfs_dmapi/xfsjunk.c - 1.1 linux/fs/Makefile - 1.43 linux/fs/xfs/Makefile - 1.130 linux/fs/xfs/linux/xfs_super.c - 1.147 - No Message Supplied linux/fs/xfs/dmapi/Makefile 1.12 renamed to linux/fs/xfs_dmapi/Makefile 1.1 - linux/fs/xfs/dmapi/Makefile 1.11 Renamed to linux/fs/xfs_dmapi/MakefileNo Message Supplied linux/fs/xfs/dmapi/dmapi_attr.c 1.5 renamed to linux/fs/xfs_dmapi/dmapi_attr.c 1.1 - linux/fs/xfs/dmapi/dmapi_attr.c 1.4 Renamed to linux/fs/xfs_dmapi/dmapi_attr.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_bulkattr.c 1.5 renamed to linux/fs/xfs_dmapi/dmapi_bulkattr.c 1.1 - linux/fs/xfs/dmapi/dmapi_bulkattr.c 1.4 Renamed to linux/fs/xfs_dmapi/dmapi_bulkattr.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_config.c 1.6 renamed to linux/fs/xfs_dmapi/dmapi_config.c 1.1 - linux/fs/xfs/dmapi/dmapi_config.c 1.5 Renamed to linux/fs/xfs_dmapi/dmapi_config.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_dmattr.c 1.5 renamed to linux/fs/xfs_dmapi/dmapi_dmattr.c 1.1 - linux/fs/xfs/dmapi/dmapi_dmattr.c 1.4 Renamed to linux/fs/xfs_dmapi/dmapi_dmattr.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_event.c 1.6 renamed to linux/fs/xfs_dmapi/dmapi_event.c 1.1 - linux/fs/xfs/dmapi/dmapi_event.c 1.5 Renamed to linux/fs/xfs_dmapi/dmapi_event.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_handle.c 1.5 renamed to linux/fs/xfs_dmapi/dmapi_handle.c 1.1 - linux/fs/xfs/dmapi/dmapi_handle.c 1.4 Renamed to linux/fs/xfs_dmapi/dmapi_handle.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_hole.c 1.5 renamed to linux/fs/xfs_dmapi/dmapi_hole.c 1.1 - linux/fs/xfs/dmapi/dmapi_hole.c 1.4 Renamed to linux/fs/xfs_dmapi/dmapi_hole.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_io.c 1.5 renamed to linux/fs/xfs_dmapi/dmapi_io.c 1.1 - linux/fs/xfs/dmapi/dmapi_io.c 1.4 Renamed to linux/fs/xfs_dmapi/dmapi_io.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_mountinfo.c 1.9 renamed to linux/fs/xfs_dmapi/dmapi_mountinfo.c 1.1 - linux/fs/xfs/dmapi/dmapi_mountinfo.c 1.8 Renamed to linux/fs/xfs_dmapi/dmapi_mountinfo.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_private.h 1.7 renamed to linux/fs/xfs_dmapi/dmapi_private.h 1.1 - linux/fs/xfs/dmapi/dmapi_private.h 1.6 Renamed to linux/fs/xfs_dmapi/dmapi_private.hNo Message Supplied linux/fs/xfs/dmapi/dmapi_region.c 1.5 renamed to linux/fs/xfs_dmapi/dmapi_region.c 1.1 - linux/fs/xfs/dmapi/dmapi_region.c 1.4 Renamed to linux/fs/xfs_dmapi/dmapi_region.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_register.c 1.12 renamed to linux/fs/xfs_dmapi/dmapi_register.c 1.1 - linux/fs/xfs/dmapi/dmapi_register.c 1.11 Renamed to linux/fs/xfs_dmapi/dmapi_register.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_right.c 1.9 renamed to linux/fs/xfs_dmapi/dmapi_right.c 1.1 - linux/fs/xfs/dmapi/dmapi_right.c 1.8 Renamed to linux/fs/xfs_dmapi/dmapi_right.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_session.c 1.10 renamed to linux/fs/xfs_dmapi/dmapi_session.c 1.1 - linux/fs/xfs/dmapi/dmapi_session.c 1.9 Renamed to linux/fs/xfs_dmapi/dmapi_session.cNo Message Supplied linux/fs/xfs/dmapi/dmapi_sysent.c 1.11 renamed to linux/fs/xfs_dmapi/dmapi_sysent.c 1.1 - linux/fs/xfs/dmapi/dmapi_sysent.c 1.10 Renamed to linux/fs/xfs_dmapi/dmapi_sysent.cNo Message Supplied linux/fs/xfs/dmapi/Status 1.4 renamed to linux/fs/xfs_dmapi/Status 1.1 - linux/fs/xfs/dmapi/Status 1.3 Renamed to linux/fs/xfs_dmapi/StatusNo Message Supplied linux/fs/Makefile.in.append - 1.5 linux/fs/xfs/Makefile.in - 1.9 - No Message Supplied linux/fs/xfs/dmapi/Makefile.in 1.2 renamed to linux/fs/xfs_dmapi/Makefile.in 1.1 - linux/fs/xfs/dmapi/Makefile.in 1.1 Renamed to linux/fs/xfs_dmapi/Makefile.inNo Message Supplied From owner-linux-xfs@oss.sgi.com Tue Dec 4 14:52:44 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB4MqiH26511 for linux-xfs-outgoing; Tue, 4 Dec 2001 14:52:44 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB4Mqfo26488 for ; Tue, 4 Dec 2001 14:52:41 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id NAA15969 for ; Tue, 4 Dec 2001 13:52:28 -0800 (PST) mail_from (roehrich@sgi.com) Received: from thistle-e185.americas.sgi.com (thistle-e185.americas.sgi.com [128.162.185.204]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id PAA3767918 for ; Tue, 4 Dec 2001 15:51:22 -0600 (CST) Received: from clink.americas.sgi.com (clink-eth.americas.sgi.com [128.162.2.8]) by thistle-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id PAA59372 for ; Tue, 4 Dec 2001 15:51:22 -0600 (CST) Received: (from roehrich@localhost) by clink.americas.sgi.com (SGI-8.9.3/8.9.3/erikj-IRIX) id PAA72301 for linux-xfs@oss.sgi.com; Tue, 4 Dec 2001 15:51:22 -0600 (CST) Date: Tue, 4 Dec 2001 15:51:22 -0600 (CST) From: Dean Roehrich Message-Id: <200112042151.PAA72301@clink.americas.sgi.com> To: linux-xfs@oss.sgi.com Subject: /proc/fs/xfs/dmapi is now /proc/fs/xfs_dmapi Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Tue Dec 4 13:50:02 PST 2001 Workarea: clink-eth.americas.sgi.com:/data/clink/a67/roehrich/2.4.x-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107765a cmd/dmapi/libdm/dmapi_lib.c - 1.6 - find the dmapi device in its new location From owner-linux-xfs@oss.sgi.com Tue Dec 4 16:38:09 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB50c9Z30259 for linux-xfs-outgoing; Tue, 4 Dec 2001 16:38:09 -0800 Received: from zok.sgi.com (zok.sgi.com [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB50c0o30237 for ; Tue, 4 Dec 2001 16:38:00 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB4H8CA10958 for ; Tue, 4 Dec 2001 09:08:12 -0800 Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3758293; Tue, 4 Dec 2001 11:06:56 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id LAA44576; Tue, 4 Dec 2001 11:06:56 -0600 (CST) Subject: Re: Is xfs_fsr safe? From: Eric Sandeen To: Walt H Cc: linux-xfs@oss.sgi.com In-Reply-To: <3C0CDF48.6020704@mindspring.com> References: <3C0CDF48.6020704@mindspring.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 04 Dec 2001 11:00:37 -0600 Message-Id: <1007485237.12151.4.camel@stout.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On the one hand, fsr is probably safe, although it doesn't get a whole lot of testing* because... ...on the other hand, it's rarely needed. In fact, XFS existed for a few years without this tool. Because of the way XFS allocates space on disk, fragmentation is kept at a minimum. You can check fragmentation on an unmounted filesystem: [root@stout root]# xfs_db /dev/sda1 xfs_db: frag actual 2204, ideal 2196, fragmentation factor 0.36% See man xfs_db for more info... -Eric *QA 042 does exercise xfs_fsr a bit, though: # XFS QA Test No. 042 # $Id: 042,v 1.2 2000/09/27 00:24:22 ajag Exp ajag $ # # xfs_fsr QA tests # create a large fragmented file and check that xfs_fsr doesn't corrupt # it or the other contents of the filesystem On Tue, 2001-12-04 at 08:35, Walt H wrote: > Hello all, > > I've been using xfs for nearly one year on one of my linux boxes, and > after some minor glitches early on, all is working well. Great work! My > question concerns xfs_fsr. Is it safe to use? Will I see any benefit if > I set it up to run as a cron job maybe twice a week? I read through the > man page and it sounds like something I should be doing for regular > maintenance, I'm just a bit timid about reorganizing a mounted > filesystem. The particular system I'm talking of is an Athlon XP 2.4.16 > kernel compiled with kgcc running 2 IBM GXP60 40GB striped. Thanks in > advance. > > -Walt -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Tue Dec 4 16:42:21 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB50gLH30480 for linux-xfs-outgoing; Tue, 4 Dec 2001 16:42:21 -0800 Received: from zok.sgi.com (zok.sgi.com [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB50gFo30458 for ; Tue, 4 Dec 2001 16:42:15 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB4JZ5A28065 for ; Tue, 4 Dec 2001 11:35:05 -0800 Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id NAA3766726; Tue, 4 Dec 2001 13:33:49 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id NAA99495; Tue, 4 Dec 2001 13:33:48 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB4JXZW19816; Tue, 4 Dec 2001 13:33:35 -0600 Subject: Re: error making real-time filesystem + logging question From: Steve Lord To: Roy Sigurd Karlsbakk Cc: "Ralf G. R. Bergs" , XFS Mailing list , stimits@idcomm.com In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 04 Dec 2001 13:33:35 -0600 Message-Id: <1007494415.13856.12.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, 2001-12-04 at 11:31, Roy Sigurd Karlsbakk wrote: > ok... > > What's the main difference between a realtime partition and a data > partiton? Can a datapartition (or non-realtime partition) be given an > extsize= parameter? That's the one I REALLY need... > On Tue, 2001-12-04 at 12:04, D. Stimits wrote: > > Now I'm curious, what's the basic purpose and performance difference > between a real-time filesystem and a "normal" filesystem? Is it some > sort of caching behavior change more suitable for streaming media? > > Just curious. > > D. Stimits, stimits@idcomm.com OK, here goes, XFS realtime 101 The XFS realtime subvolume is a separate dataspace in the filesystem, it is only used for files marked as realtime after they are created. There is nothing in the realtime subvolume except file data, where as in the normal XFS data subvolume there is a mix of metadata and file data. Space allocation in the realtime subvolume is managed by a different allocator which gives out space in multiples of a pre-specified size, the allocator uses a binary chop approach to space allocation which is designed to avoid fragmentation at the expense of efficient use of space - the allocator should also be faster. The end result is that streaming I/O should behave better on realtime files and on regular files. On Irix the realtime subvolume is sometimes used in combination with GRIO - which is a bandwidth guarantee subsystem, an application can request X amount of bandwidth for Y amount of time, the system does its best to schedule I/O to guarantee these requests are met. There is no plan to offer GRIO on Linux at this time, and the features of the realtime allocator are not available for regular files. One feature which can be used on normal files which might help with placement is preallocation of space. Take a look at the use of the XFS_IOC_RESVSP64 ioctl in cmd/xfsprogs/mkfile/xfs_mkfile.c Steve -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Tue Dec 4 17:01:07 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5117331673 for linux-xfs-outgoing; Tue, 4 Dec 2001 17:01:07 -0800 Received: from lips.thebarn.com (lips.borg.umn.edu [160.94.232.50]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5113o31648 for ; Tue, 4 Dec 2001 17:01:03 -0800 Received: from scare.vieo.com ([63.231.179.33]) by lips.thebarn.com (8.12.0/8.12.0) with ESMTP id fB4NxlML013110 for ; Tue, 4 Dec 2001 17:59:48 -0600 (CST) Subject: html filter test From: Russell Cattelan To: linux-xfs@oss.sgi.com X-Mailer: Evolution/0.99.0 (Preview Release) Date: 04 Dec 2001 17:53:24 -0600 Message-Id: <1007510005.2534.2.camel@scare> Mime-Version: 1.0 Content-Disposition: inline Content-Type: text/plain Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk This is html mail ... but hopefully it will look like normal mail Fun color stuff here. More .... [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Tue Dec 4 17:26:44 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB51QiQ32288 for linux-xfs-outgoing; Tue, 4 Dec 2001 17:26:44 -0800 Received: from lips.thebarn.com (lips.borg.umn.edu [160.94.232.50]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB51Qao32262 for ; Tue, 4 Dec 2001 17:26:36 -0800 Received: from scare.vieo.com ([63.231.179.33]) by lips.thebarn.com (8.12.0/8.12.0) with ESMTP id fB50PKML013401 for ; Tue, 4 Dec 2001 18:25:21 -0600 (CST) Subject: Re: html filter test From: Russell Cattelan To: linux-xfs@oss.sgi.com In-Reply-To: <20011205000427.GG1152@gwyn.tux.org> References: <1007510005.2534.2.camel@scare> <20011205000427.GG1152@gwyn.tux.org> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.0 (Preview Release) Date: 04 Dec 2001 18:18:57 -0600 Message-Id: <1007511537.2534.11.camel@scare> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, 2001-12-04 at 18:04, Timothy Ball wrote: > Hey how'd you do that? I've been meaning to put something like this for > my lug list. > Since 3 people asked about this already I'm posting this to the whole list. Original article http://www.unixreview.com/documents/s=1236/urm0104h/0104h.htm script all typed up http://www.kluge.net/mailfiltering/strip_html.txt Also linux-xfs xfs was already running through procmail... /etc/aliases: #linux-xfs: "| wrapper resend -l linux-xfs linux-xfs-outgoing" linux-xfs: "| procmail -m LIST='linux-xfs' /etc/mail/procmailrc.lists" etc.... procmail.lists: #VERBOSE=on MAILDIR=$HOME/Mail LOGFILE=majorlog PATH=/usr/bin:/usr/local/bin:/bin HOME=/etc/mail/prochome/ MAILDIR=$HOME LOGFILE=$MAILDIR/majorlog-`date +%y%m` JFDIR=/etc/mail/junkfilter/ :0 Whc: msgid.lock | formail -D 8192 msgid.cache :0 a: duplicates-`date +%y%m` # Call junkfilter INCLUDERC=$JFDIR/junkfilter # Take action if junkfilter caught a junkmail. :0 * JFEXP ?? . { :0 f * JFSTATUS ?? 1 | formail -i "X-junkfilter: $JFVERSION" -i "X-Spammer: $JFEXP" :0 E : | formail -i "X-junkfilter: $JFVERSION" -i "X-Spammer: $JFEXP" \ >> junkmail-`date +%y%m` } :0 fw * ^Content-type:.*boundary * <100000 |$HOME/strip_html.pl :0 | /etc/smrsh/wrapper resend -l $LIST $LIST-outgoing > -- > GPG key available on pgpkeys.mit.edu > pub 1024D/511FBD54 2001-07-23 Timothy Lu Hu Ball > Key fingerprint = B579 29B0 F6C8 C7AA 3840 E053 FE02 BB97 511F BD54 From owner-linux-xfs@oss.sgi.com Tue Dec 4 18:15:15 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB52FFe00657 for linux-xfs-outgoing; Tue, 4 Dec 2001 18:15:15 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB52FAo00634 for ; Tue, 4 Dec 2001 18:15:11 -0800 Received: from nodin.corp.sgi.com (nodin.corp.sgi.com [192.26.51.193]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id CAA1430177 for ; Wed, 5 Dec 2001 02:15:05 +0100 (CET) mail_from (kaos@melbourne.sgi.com) Received: from kao2.melbourne.sgi.com (kao2.melbourne.sgi.com [134.14.55.180]) by nodin.corp.sgi.com (8.11.4/8.11.2/nodin-1.0) with ESMTP id fB51E3416246525; Tue, 4 Dec 2001 17:14:03 -0800 (PST) Received: by kao2.melbourne.sgi.com (Postfix, from userid 16331) id DB8EC300090; Wed, 5 Dec 2001 12:14:01 +1100 (EST) Received: from kao2.melbourne.sgi.com (localhost [127.0.0.1]) by kao2.melbourne.sgi.com (Postfix) with ESMTP id 5FE4196; Wed, 5 Dec 2001 12:14:01 +1100 (EST) X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 From: Keith Owens To: Dean Roehrich Cc: linux-xfs@oss.sgi.com Subject: Re: TAKE - switch CONFIG_XFS_DMAPI to CONFIG_HAVE_XFS_DMAPI, prep for module In-reply-to: Your message of "Tue, 04 Dec 2001 14:03:20 MDT." <200112042003.OAA54282@clink.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Wed, 05 Dec 2001 12:13:55 +1100 Message-ID: <2181.1007514835@kao2.melbourne.sgi.com> Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, 4 Dec 2001 14:03:20 -0600 (CST), Dean Roehrich wrote: >If I build dmapi as a module then CONFIG_XFS_DMAPI won't be set--it'll >be CONFIG_XFS_DMAPI_MODULE instead. So switch to using CONFIG_HAVE_XFS_DMAPI >to see if DMAPI is being used, whether dmapi is a module or not. Should that variable have _XFS_ in the name or should there be a standard name that applies to all dmapi implementations? IOW, is dmapi XFS local or part of a wider filesystem standard? From owner-linux-xfs@oss.sgi.com Tue Dec 4 18:52:09 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB52q9K01537 for linux-xfs-outgoing; Tue, 4 Dec 2001 18:52:09 -0800 Received: from zok.sgi.com (zok.sgi.com [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB52q5o01515 for ; Tue, 4 Dec 2001 18:52:05 -0800 Received: from sherman.melbourne.sgi.com (sherman.melbourne.sgi.com [134.14.55.175]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB51pvA20203 for ; Tue, 4 Dec 2001 17:51:58 -0800 Received: (from kaos@localhost) by sherman.melbourne.sgi.com (8.9.3/8.9.3) id MAA26858; Wed, 5 Dec 2001 12:51:33 +1100 Date: Wed, 5 Dec 2001 12:51:33 +1100 From: Keith Owens Message-Id: <200112050151.MAA26858@sherman.melbourne.sgi.com> Subject: TAKE - Reorder xfs Makefile.in files Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk The preferred order for Makefile.in entries is Global structure (expsyms). Local structure (objlink). Selection (select, link_subdirs). Flags (extra_[ca]flags). Special rules (user_command). I converted from kbuild 2.4 to 2.5 using the original order for ease of debugging, time to clean up the 2.5 order. Date: Tue Dec 4 17:45:29 PST 2001 Workarea: sherman.melbourne.sgi.com:/build/kaos/2.4.x-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107788a linux/fs/pagebuf/Makefile.in - 1.2 linux/fs/xfs/Makefile.in - 1.10 linux/fs/xfs_support/Makefile.in - 1.3 linux/kdb/modules/Makefile.in - 1.5 linux/fs/xfs_dmapi/Makefile.in - 1.2 From owner-linux-xfs@oss.sgi.com Tue Dec 4 20:34:43 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB54Yhp10541 for linux-xfs-outgoing; Tue, 4 Dec 2001 20:34:43 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB54Xno10507 for ; Tue, 4 Dec 2001 20:33:49 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via SMTP id TAA18810 for ; Tue, 4 Dec 2001 19:33:33 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id OAA06608; Wed, 5 Dec 2001 14:32:14 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id OAA48082; Wed, 5 Dec 2001 14:32:10 +1100 (AEDT) Date: Wed, 5 Dec 2001 14:32:10 +1100 From: Nathan Scott To: Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: [PATCH] Revised extended attributes interface Message-ID: <20011205143209.C44610@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk > On December 3, 2001 01:54 am, Nathan Scott wrote: > > ...BTW, we have reworked the interfaces once more and will > > send out the latest revision in the next couple of days - hi folks, Here is the revised interface. I believe it takes into account the issues raised so far - further suggestions are also welcome, of course. Man pages for the system calls are available from the XFS CVS tree http://oss.sgi.com/cgi-bin/cvsweb.cgi/linux-2.4-xfs/cmd/attr2/man/ [Andreas, could you host html-ised versions at bestbits.at again?] The interesting pages are getxattr(2), setxattr(2), listxattr(2), removexattr(2) and attr(5), though there are several user tools based on this interface too and a version of Andreas' POSIX ACL tools which makes use of this interface now exists (these also have man pages and are all available from the XFS CVS tree). Two patches follow - the first marks syscall numbers as reserved, the second is the proposed VFS interface. These are patches based on the 2.5.0 tree, but should apply cleanly to any 2.5.1-preX and 2.4.16/17-preX tree. Linus - if possible, we'd really like to get system call numbers reserved for these, or know of any aspects you would like changed in order to make this acceptable for 2.5. many thanks. -- Nathan [1st patch] diff -Naur 2.5.0-pristine/arch/i386/kernel/entry.S 2.5.0-reserved/arch/i386/kernel/entry.S --- 2.5.0-pristine/arch/i386/kernel/entry.S Sat Nov 3 12:18:49 2001 +++ 2.5.0-reserved/arch/i386/kernel/entry.S Tue Dec 4 11:57:32 2001 @@ -622,6 +622,18 @@ .long SYMBOL_NAME(sys_ni_syscall) /* Reserved for Security */ .long SYMBOL_NAME(sys_gettid) .long SYMBOL_NAME(sys_readahead) /* 225 */ + .long SYMBOL_NAME(sys_ni_syscall) /* reserved for setxattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* reserved for lsetxattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* reserved for fsetxattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* reserved for getxattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* 230 reserved for lgetxattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* reserved for fgetxattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* reserved for listxattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* reserved for llistxattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* reserved for flistxattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* 235 reserved for removexattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* reserved for lremovexattr */ + .long SYMBOL_NAME(sys_ni_syscall) /* reserved for fremovexattr */ .rept NR_syscalls-(.-sys_call_table)/4 .long SYMBOL_NAME(sys_ni_syscall) diff -Naur 2.5.0-pristine/include/asm-i386/unistd.h 2.5.0-reserved/include/asm-i386/unistd.h --- 2.5.0-pristine/include/asm-i386/unistd.h Thu Oct 18 03:03:03 2001 +++ 2.5.0-reserved/include/asm-i386/unistd.h Tue Dec 4 11:58:21 2001 @@ -230,6 +230,18 @@ #define __NR_security 223 /* syscall for security modules */ #define __NR_gettid 224 #define __NR_readahead 225 +#define __NR_setxattr 226 +#define __NR_lsetxattr 227 +#define __NR_fsetxattr 228 +#define __NR_getxattr 229 +#define __NR_lgetxattr 230 +#define __NR_fgetxattr 231 +#define __NR_listxattr 232 +#define __NR_llistxattr 233 +#define __NR_flistxattr 234 +#define __NR_removexattr 235 +#define __NR_lremovexattr 236 +#define __NR_fremovexattr 237 /* user-visible error numbers are in the range -1 - -124: see */ [2nd patch] diff -Naur 2.5.0-pristine/arch/i386/kernel/entry.S 2.5.0-xattr/arch/i386/kernel/entry.S --- 2.5.0-pristine/arch/i386/kernel/entry.S Sat Nov 3 12:18:49 2001 +++ 2.5.0-xattr/arch/i386/kernel/entry.S Tue Dec 4 12:02:56 2001 @@ -622,6 +622,18 @@ .long SYMBOL_NAME(sys_ni_syscall) /* Reserved for Security */ .long SYMBOL_NAME(sys_gettid) .long SYMBOL_NAME(sys_readahead) /* 225 */ + .long SYMBOL_NAME(sys_setxattr) + .long SYMBOL_NAME(sys_lsetxattr) + .long SYMBOL_NAME(sys_fsetxattr) + .long SYMBOL_NAME(sys_getxattr) + .long SYMBOL_NAME(sys_lgetxattr) /* 230 */ + .long SYMBOL_NAME(sys_fgetxattr) + .long SYMBOL_NAME(sys_listxattr) + .long SYMBOL_NAME(sys_llistxattr) + .long SYMBOL_NAME(sys_flistxattr) + .long SYMBOL_NAME(sys_removexattr) /* 235 */ + .long SYMBOL_NAME(sys_lremovexattr) + .long SYMBOL_NAME(sys_fremovexattr) .rept NR_syscalls-(.-sys_call_table)/4 .long SYMBOL_NAME(sys_ni_syscall) diff -Naur 2.5.0-pristine/fs/Makefile 2.5.0-xattr/fs/Makefile --- 2.5.0-pristine/fs/Makefile Tue Nov 13 04:34:16 2001 +++ 2.5.0-xattr/fs/Makefile Fri Nov 30 15:33:28 2001 @@ -14,7 +14,7 @@ super.o block_dev.o char_dev.o stat.o exec.o pipe.o namei.o \ fcntl.o ioctl.o readdir.o select.o fifo.o locks.o \ dcache.o inode.o attr.o bad_inode.o file.o iobuf.o dnotify.o \ - filesystems.o namespace.o seq_file.o + filesystems.o namespace.o seq_file.o xattr.o ifeq ($(CONFIG_QUOTA),y) obj-y += dquot.o diff -Naur 2.5.0-pristine/fs/xattr.c 2.5.0-xattr/fs/xattr.c --- 2.5.0-pristine/fs/xattr.c Thu Jan 1 10:00:00 1970 +++ 2.5.0-xattr/fs/xattr.c Tue Dec 4 12:00:49 2001 @@ -0,0 +1,346 @@ +/* + File: fs/xattr.c + + Extended attribute handling. + + Copyright (C) 2001 by Andreas Gruenbacher + Copyright (C) 2001 SGI - Silicon Graphics, Inc + */ +#include +#include +#include +#include +#include +#include +#include + +/* + * Extended attribute memory allocation wrappers, originally + * based on the Intermezzo PRESTO_ALLOC/PRESTO_FREE macros. + * The vmalloc use here is very uncommon - extended attributes + * are supposed to be small chunks of metadata, and it is quite + * unusual to have very many extended attributes, so lists tend + * to be quite short as well. The 64K upper limit is derived + * from the extended attribute size limit used by XFS. + * Intentionally allow zero @size for value/list size requests. + */ +static void * +xattr_alloc(size_t size, size_t limit) +{ + void *ptr; + + if (size > limit) + return ERR_PTR(-E2BIG); + + if (!size) /* size request, no buffer is needed */ + return NULL; + else if (size <= PAGE_SIZE) + ptr = kmalloc((unsigned long) size, GFP_KERNEL); + else + ptr = vmalloc((unsigned long) size); + if (!ptr) + return ERR_PTR(-ENOMEM); + return ptr; +} + +static void +xattr_free(void *ptr, size_t size) +{ + if (!size) /* size request, no buffer was needed */ + return; + else if (size <= PAGE_SIZE) + kfree(ptr); + else + vfree(ptr); +} + +/* + * Extended attribute SET operations + */ +static long +setxattr(struct dentry *d, char *name, void *value, size_t size, int flags) +{ + int error; + void *kvalue; + char kname[XATTR_NAME_MAX + 1]; + + error = -EINVAL; + if (flags & ~(XATTR_CREATE|XATTR_REPLACE)) + return error; + + error = -EFAULT; + if (copy_from_user(kname, name, XATTR_NAME_MAX)) + return error; + kname[XATTR_NAME_MAX] = '\0'; + + kvalue = xattr_alloc(size, XATTR_SIZE_MAX); + if (IS_ERR(kvalue)) + return PTR_ERR(kvalue); + + error = -EFAULT; + if (size > 0 && copy_from_user(kvalue, value, size)) { + xattr_free(kvalue, size); + return error; + } + + error = -EOPNOTSUPP; + if (d->d_inode->i_op && d->d_inode->i_op->setxattr) { + lock_kernel(); + error = d->d_inode->i_op->setxattr(d, kname, kvalue, size, flags); + unlock_kernel(); + } + + xattr_free(kvalue, size); + return error; +} + +asmlinkage long +sys_setxattr(char *path, char *name, void *value, size_t size, int flags) +{ + struct nameidata nd; + int error; + + error = user_path_walk(path, &nd); + if (error) + return error; + error = setxattr(nd.dentry, name, value, size, flags); + path_release(&nd); + return error; +} + +asmlinkage long +sys_lsetxattr(char *path, char *name, void *value, size_t size, int flags) +{ + struct nameidata nd; + int error; + + error = user_path_walk_link(path, &nd); + if (error) + return error; + error = setxattr(nd.dentry, name, value, size, flags); + path_release(&nd); + return error; +} + +asmlinkage long +sys_fsetxattr(int fd, char *name, void *value, size_t size, int flags) +{ + struct file *f; + int error = -EBADF; + + f = fget(fd); + if (!f) + return error; + error = setxattr(f->f_dentry, name, value, size, flags); + fput(f); + return error; +} + +/* + * Extended attribute GET operations + */ +static long +getxattr(struct dentry *d, char *name, void *value, size_t size) +{ + int error; + void *kvalue; + char kname[XATTR_NAME_MAX + 1]; + + error = -EFAULT; + if (copy_from_user(kname, name, XATTR_NAME_MAX)) + return error; + kname[XATTR_NAME_MAX] = '\0'; + + kvalue = xattr_alloc(size, XATTR_SIZE_MAX); + if (IS_ERR(kvalue)) + return PTR_ERR(kvalue); + + error = -EOPNOTSUPP; + if (d->d_inode->i_op && d->d_inode->i_op->getxattr) { + lock_kernel(); + error = d->d_inode->i_op->getxattr(d, kname, kvalue, size); + unlock_kernel(); + } + + if (kvalue && error > 0) + if (copy_to_user(value, kvalue, size)) + error = -EFAULT; + xattr_free(kvalue, size); + return error; +} + +asmlinkage long +sys_getxattr(char *path, char *name, void *value, size_t size) +{ + struct nameidata nd; + int error; + + error = user_path_walk(path, &nd); + if (error) + return error; + error = getxattr(nd.dentry, name, value, size); + path_release(&nd); + return error; +} + +asmlinkage long +sys_lgetxattr(char *path, char *name, void *value, size_t size) +{ + struct nameidata nd; + int error; + + error = user_path_walk_link(path, &nd); + if (error) + return error; + error = getxattr(nd.dentry, name, value, size); + path_release(&nd); + return error; +} + +asmlinkage long +sys_fgetxattr(int fd, char *name, void *value, size_t size) +{ + struct file *f; + int error = -EBADF; + + f = fget(fd); + if (!f) + return error; + error = getxattr(f->f_dentry, name, value, size); + fput(f); + return error; +} + +/* + * Extended attribute LIST operations + */ +static long +listxattr(struct dentry *d, char *list, size_t size) +{ + int error; + char *klist; + + klist = (char *)xattr_alloc(size, XATTR_LIST_MAX); + if (IS_ERR(klist)) + return PTR_ERR(klist); + + error = -EOPNOTSUPP; + if (d->d_inode->i_op && d->d_inode->i_op->listxattr) { + lock_kernel(); + error = d->d_inode->i_op->listxattr(d, klist, size); + unlock_kernel(); + } + + if (klist && error > 0) + if (copy_to_user(list, klist, size)) + error = -EFAULT; + xattr_free(klist, size); + return error; +} + +asmlinkage long +sys_listxattr(char *path, char *list, size_t size) +{ + struct nameidata nd; + int error; + + error = user_path_walk(path, &nd); + if (error) + return error; + error = listxattr(nd.dentry, list, size); + path_release(&nd); + return error; +} + +asmlinkage long +sys_llistxattr(char *path, char *list, size_t size) +{ + struct nameidata nd; + int error; + + error = user_path_walk_link(path, &nd); + if (error) + return error; + error = listxattr(nd.dentry, list, size); + path_release(&nd); + return error; +} + +asmlinkage long +sys_flistxattr(int fd, char *list, size_t size) +{ + struct file *f; + int error = -EBADF; + + f = fget(fd); + if (!f) + return error; + error = listxattr(f->f_dentry, list, size); + fput(f); + return error; +} + +/* + * Extended attribute REMOVE operations + */ +static long +removexattr(struct dentry *d, char *name) +{ + int error; + char kname[XATTR_NAME_MAX + 1]; + + error = -EFAULT; + if (copy_from_user(kname, name, XATTR_NAME_MAX)) + return error; + kname[XATTR_NAME_MAX] = '\0'; + + error = -EOPNOTSUPP; + if (d->d_inode->i_op && d->d_inode->i_op->removexattr) { + lock_kernel(); + error = d->d_inode->i_op->removexattr(d, kname); + unlock_kernel(); + } + return error; +} + +asmlinkage long +sys_removexattr(char *path, char *name) +{ + struct nameidata nd; + int error; + + error = user_path_walk(path, &nd); + if (error) + return error; + error = removexattr(nd.dentry, name); + path_release(&nd); + return error; +} + +asmlinkage long +sys_lremovexattr(char *path, char *name) +{ + struct nameidata nd; + int error; + + error = user_path_walk_link(path, &nd); + if (error) + return error; + error = removexattr(nd.dentry, name); + path_release(&nd); + return error; +} + +asmlinkage long +sys_fremovexattr(int fd, char *name) +{ + struct file *f; + int error = -EBADF; + + f = fget(fd); + if (!f) + return error; + error = removexattr(f->f_dentry, name); + fput(f); + return error; +} diff -Naur 2.5.0-pristine/include/asm-i386/unistd.h 2.5.0-xattr/include/asm-i386/unistd.h --- 2.5.0-pristine/include/asm-i386/unistd.h Thu Oct 18 03:03:03 2001 +++ 2.5.0-xattr/include/asm-i386/unistd.h Tue Dec 4 12:03:22 2001 @@ -230,6 +230,18 @@ #define __NR_security 223 /* syscall for security modules */ #define __NR_gettid 224 #define __NR_readahead 225 +#define __NR_setxattr 226 +#define __NR_lsetxattr 227 +#define __NR_fsetxattr 228 +#define __NR_getxattr 229 +#define __NR_lgetxattr 230 +#define __NR_fgetxattr 231 +#define __NR_listxattr 232 +#define __NR_llistxattr 233 +#define __NR_flistxattr 234 +#define __NR_removexattr 235 +#define __NR_lremovexattr 236 +#define __NR_fremovexattr 237 /* user-visible error numbers are in the range -1 - -124: see */ diff -Naur 2.5.0-pristine/include/linux/fs.h 2.5.0-xattr/include/linux/fs.h --- 2.5.0-pristine/include/linux/fs.h Fri Nov 23 06:46:19 2001 +++ 2.5.0-xattr/include/linux/fs.h Tue Dec 4 12:03:34 2001 @@ -851,6 +851,10 @@ int (*revalidate) (struct dentry *); int (*setattr) (struct dentry *, struct iattr *); int (*getattr) (struct dentry *, struct iattr *); + int (*setxattr) (struct dentry *, char *, void *, size_t, int); + int (*getxattr) (struct dentry *, char *, void *, size_t); + int (*listxattr) (struct dentry *, char *, size_t); + int (*removexattr) (struct dentry *, char *); }; /* diff -Naur 2.5.0-pristine/include/linux/limits.h 2.5.0-xattr/include/linux/limits.h --- 2.5.0-pristine/include/linux/limits.h Thu Jul 29 03:30:10 1999 +++ 2.5.0-xattr/include/linux/limits.h Fri Nov 30 15:33:28 2001 @@ -13,6 +13,9 @@ #define NAME_MAX 255 /* # chars in a file name */ #define PATH_MAX 4095 /* # chars in a path name */ #define PIPE_BUF 4096 /* # bytes in atomic write to a pipe */ +#define XATTR_NAME_MAX 255 /* # chars in an extended attribute name */ +#define XATTR_SIZE_MAX 65536 /* size of an extended attribute value (64k) */ +#define XATTR_LIST_MAX 65536 /* size of extended attribute namelist (64k) */ #define RTSIG_MAX 32 diff -Naur 2.5.0-pristine/include/linux/xattr.h 2.5.0-xattr/include/linux/xattr.h --- 2.5.0-pristine/include/linux/xattr.h Thu Jan 1 10:00:00 1970 +++ 2.5.0-xattr/include/linux/xattr.h Tue Dec 4 12:01:35 2001 @@ -0,0 +1,15 @@ +/* + File: linux/xattr.h + + Extended attributes handling. + + Copyright (C) 2001 by Andreas Gruenbacher + Copyright (C) 2001 SGI - Silicon Graphics, Inc +*/ +#ifndef _LINUX_XATTR_H +#define _LINUX_XATTR_H + +#define XATTR_CREATE 0x1 /* set value, fail if attr already exists */ +#define XATTR_REPLACE 0x2 /* set value, fail if attr does not exist */ + +#endif /* _LINUX_XATTR_H */ From owner-linux-xfs@oss.sgi.com Tue Dec 4 22:18:09 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB56I9F12301 for linux-xfs-outgoing; Tue, 4 Dec 2001 22:18:09 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB56I3o12278 for ; Tue, 4 Dec 2001 22:18:03 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via SMTP id VAA05251 for ; Tue, 4 Dec 2001 21:18:02 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id QAA07155; Wed, 5 Dec 2001 16:16:42 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id QAA45952; Wed, 5 Dec 2001 16:16:40 +1100 (AEDT) Date: Wed, 5 Dec 2001 16:16:40 +1100 From: Nathan Scott To: Roy Sigurd Karlsbakk Cc: XFS Mailing list Subject: Re: error making real-time filesystem + logging question Message-ID: <20011205161640.E44610@wobbly.melbourne.sgi.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from roy@karlsbakk.net on Tue, Dec 04, 2001 at 05:45:35PM +0100 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi Roy, On Tue, Dec 04, 2001 at 05:45:35PM +0100, Roy Sigurd Karlsbakk wrote: > hi all > > I'm trying to make a real-time filesystem for testing purposes onto a > single disk. (yes - I know it's silly, but it's only for testing). > > I'm trying to do this on one drive with two partitions, and it all messes > up. see below for more info. > > roy > > -- > from /proc/partitions > major minor #blocks name > 3 69 19277937 hdb5 > 3 70 730926 hdb6 > -- > # mkfs.xfs -f -r extsize=4m,rtdev=/dev/hdb5 /dev/hdb6 > meta-data=/dev/hdb6 isize=256 agcount=8, agsize=22842 blks > data = bsize=4096 blocks=182731, imaxpct=25 > = sunit=0 swidth=0 blks, unwritten=0 > = imaxbits=0 > naming =version 2 bsize=4096 > log =internal log bsize=4096 blocks=1200 > realtime =/dev/hdb5 extsz=4194304 blocks=4819484, > rtextents=4706 > mkfs.xfs: lseek64 to 19740605952 failed: Invalid argument > > but... > > # mkfs.xfs -f -r extsize=4m,rtdev=/dev/hdb6 /dev/hdb5 > > works fine... ? que? > This is a bug in Linux mkfs.xfs where it is checking the size of the device - the code is currently incorrecly inspecting the data device instead of the realtime device here, so the lseek fails if the data device is smaller than the realtime device. I'll get a fix in shortly -- thanks for reporting the problem. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Tue Dec 4 22:50:35 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB56oZQ12885 for linux-xfs-outgoing; Tue, 4 Dec 2001 22:50:35 -0800 Received: from mail-nb00s0.nbnet.nb.ca (smtp1.nbnet.nb.ca [198.164.200.23]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB56oWo12853 for ; Tue, 4 Dec 2001 22:50:32 -0800 Received: from wizard ([156.34.217.173]) by mail-nb00s0.nbnet.nb.ca (Post.Office MTA v3.5.3 release 223 ID# 0-72041U145000L145000S0V35) with SMTP id ca for ; Wed, 5 Dec 2001 01:50:30 -0400 Reply-To: From: "David Wasylciw" To: Subject: Problems setting ACL with setfacl Date: Wed, 5 Dec 2001 01:50:51 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Importance: Normal Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hello All, I've recently installed kernel 2.4.16 with the latest XFS all together patch and of course compiled in ACL support. I'm running Redhat 6.2 and have installed all the latest acl programs and libraries but for some reason when I attempt to add a user to the ACL list for a file I get the following error: setfacl: file: Resulting ACL `,user::rw-,user:dwasyl:rw-,mask::---,other::r--': Invalid entry type at entry 1 This was generated with the command: setfacl -m u:dwasyl:rw testfile I checked through the archives and saw something about perhaps needing to have the permissions listed as rw- but that did not seem to help and generated the exact same error. It would be much appreciated if anyone could help point me in the right direction and if any more information is needed just let me know. Thanks, Dave From owner-linux-xfs@oss.sgi.com Tue Dec 4 23:27:04 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB57R4313531 for linux-xfs-outgoing; Tue, 4 Dec 2001 23:27:04 -0800 Received: from mail-nb00s0.nbnet.nb.ca (smtp1.nbnet.nb.ca [198.164.200.23]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB57R1o13509 for ; Tue, 4 Dec 2001 23:27:01 -0800 Received: from wizard ([156.34.217.173]) by mail-nb00s0.nbnet.nb.ca (Post.Office MTA v3.5.3 release 223 ID# 0-72041U145000L145000S0V35) with SMTP id ca for ; Wed, 5 Dec 2001 02:26:59 -0400 Reply-To: From: "David Wasylciw" To: Subject: Re: Problems setting ACL with setfacl Date: Wed, 5 Dec 2001 02:27:19 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) Importance: Normal X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Alright, well I found some more posts and discovered that I just shouldn't be using setfacl at the moment but not using that how would I go about setting the ACLs? That's the only program I can find with documentation. Thanks, sorry about the prior question. Dave From owner-linux-xfs@oss.sgi.com Tue Dec 4 23:33:33 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB57XXs13721 for linux-xfs-outgoing; Tue, 4 Dec 2001 23:33:33 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB57XTo13699 for ; Tue, 4 Dec 2001 23:33:29 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via SMTP id HAA1455991 for ; Wed, 5 Dec 2001 07:33:25 +0100 (CET) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id RAA07632; Wed, 5 Dec 2001 17:32:03 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id RAA48552; Wed, 5 Dec 2001 17:32:02 +1100 (AEDT) Date: Wed, 5 Dec 2001 17:32:02 +1100 From: Nathan Scott To: David Wasylciw Cc: linux-xfs@oss.sgi.com Subject: Re: Problems setting ACL with setfacl Message-ID: <20011205173201.I44610@wobbly.melbourne.sgi.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from david@wasylciw.com on Wed, Dec 05, 2001 at 02:27:19AM -0400 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi, On Wed, Dec 05, 2001 at 02:27:19AM -0400, David Wasylciw wrote: > Alright, well I found some more posts and discovered that I just shouldn't > be using setfacl at the moment but not using that how would I go about > setting the ACLs? That's the only program I can find with documentation. > No, you were right originally - it should work. I'm about to head off for the day - I'll have a more detailed look tomorrow. In the meantime, try the chacl(1) command. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Tue Dec 4 23:53:41 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB57rfR14504 for linux-xfs-outgoing; Tue, 4 Dec 2001 23:53:41 -0800 Received: from mta03ps.bigpond.com (mta03ps.bigpond.com [144.135.25.135]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB57rbo14480 for ; Tue, 4 Dec 2001 23:53:37 -0800 Message-Id: <200112050753.fB57rbo14480@oss.sgi.com> Received: from there ([144.135.25.78]) by mta03ps.bigpond.com (Netscape Messaging Server 4.15) with SMTP id GNUZGC00.9UQ for ; Wed, 5 Dec 2001 17:00:12 +1000 Received: from 144.137.136.128 ([144.137.136.128]) by PSMAM04.mailsvc.email.bigpond.com(MailRouter V2.9k 8407/38868540); 05 Dec 2001 16:53:21 Content-Type: text/plain; charset="us-ascii" From: Adrian Head To: linux-xfs@oss.sgi.com Subject: How do you calculate optimal amount of RAM required for XFS filesystems? Date: Wed, 5 Dec 2001 16:53:18 +1000 X-Mailer: KMail [version 1.3.1] MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk I have been lurking on this list quick some time now; but I have never seen a discussion regarding how to calculate the optimum amount of RAM required for XFS based filesystems. There was a thread a very long time ago (cannot find it now) where it was stated that leaps of RAM was required but didn't specify the quantity of heaps ;-) For memory it also discussed the caching of filesystem data for performance. But I don't think the interaction between amount of RAM and XFS was discussed fully. Does anyone have experience on this subject? Say - how much RAM is required for a 80G or 120G XFS file system? Does having many different sized XFS partitions change anything? For example I have 1x 40G, 1x80G, 1x120G XFS partitions on the same fileserver with 384M RAM. Is this enough? -- Adrian Head (Public Key available on request.) From owner-linux-xfs@oss.sgi.com Wed Dec 5 00:13:45 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB58Djg21944 for linux-xfs-outgoing; Wed, 5 Dec 2001 00:13:45 -0800 Received: from zok.sgi.com (zok.SGI.COM [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB58Dfo21922 for ; Wed, 5 Dec 2001 00:13:41 -0800 Received: from snort.melbourne.sgi.com (snort.melbourne.sgi.com [134.14.55.149]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB57DYA00478 for ; Tue, 4 Dec 2001 23:13:34 -0800 Received: (from nathans@localhost) by snort.melbourne.sgi.com (SGI-8.9.3/8.9.3) id SAA42521; Wed, 5 Dec 2001 18:12:16 +1100 (EST) Date: Wed, 5 Dec 2001 18:12:16 +1100 (EST) From: Nathan Scott Message-Id: <200112050712.SAA42521@snort.melbourne.sgi.com> To: linux-xfs@oss.sgi.com Cc: roy@karlsbakk.net Subject: TAKE - xfsprogs Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Tue Dec 4 23:11:19 PST 2001 Workarea: snort.melbourne.sgi.com:/hosts/snort/home/diskb/build4/nathans/linux-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107799a cmd/xfsprogs/growfs/xfs_growfs.c - 1.7 cmd/xfsprogs/mkfs/xfs_mkfs.c - 1.18 cmd/xfsprogs/man/man8/xfs_growfs.8 - 1.3 - remove remaining inodebits calculation code, negated by recent changes in the XFS kernel code. cmd/xfsprogs/libxfs/xfs_rtalloc.c - 1.5 cmd/xfsprogs/libxfs/xfs.h - 1.9 - fix a bug in the validation of the size of the realtime device. cmd/xfsprogs/include/xfs_mount.h - 1.7 - sync with kernel header. cmd/xfsprogs/doc/CHANGES - 1.45 cmd/xfsprogs/debian/changelog - 1.33 - document changes since 1.3.13. cmd/xfsprogs/VERSION - 1.35 - bump version to 1.3.14. From owner-linux-xfs@oss.sgi.com Wed Dec 5 00:50:54 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB58osq24336 for linux-xfs-outgoing; Wed, 5 Dec 2001 00:50:54 -0800 Received: from smtpzilla5.xs4all.nl (smtpzilla5.xs4all.nl [194.109.127.141]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB58oko24314 for ; Wed, 5 Dec 2001 00:50:46 -0800 Received: from auto-nb1.xs4all.nl (coltex.xs4all.nl [213.84.127.168]) by smtpzilla5.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB57oeKf069603; Wed, 5 Dec 2001 08:50:40 +0100 (CET) Message-Id: <4.3.2.7.2.20011205083836.035ca7b0@pop.xs4all.nl> X-Sender: knuffie@pop.xs4all.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Wed, 05 Dec 2001 08:47:34 +0100 To: Adrian Head , linux-xfs@oss.sgi.com From: Seth Mos Subject: Re: How do you calculate optimal amount of RAM required for XFS filesystems? In-Reply-To: <200112050753.fB57rbo14480@oss.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 16:53 5-12-2001 +1000, Adrian Head wrote: >I have been lurking on this list quick some time now; but I have never seen a >discussion regarding how to calculate the optimum amount of RAM required for >XFS based filesystems. > >There was a thread a very long time ago (cannot find it now) where it was >stated that leaps of RAM was required but didn't specify the quantity of >heaps ;-) For memory it also discussed the caching of filesystem data for >performance. But I don't think the interaction between amount of RAM and XFS >was discussed fully. Some of the older code needed more ram because the relied on getting it when it asked for it (during recovery for example). These days 12MB is a minimum requirement to boot a XFS box and perform recovery as well. If you have more then this you just start caching data and everything feels faster. If you have a webserver you could use 256 MB or more to make sure that the speed stays up by caching a lot of the static data. If your working static working set is about 1 GB of data, 512 MB ram should be able to cache a lot of the requests. Also take into account that certain programs run a _lot_ faster when they can also fit in memory. >Does anyone have experience on this subject? > >Say - how much RAM is required for a 80G or 120G XFS file system? 32MB would work just fine and more is better. >Does having many different sized XFS partitions change anything? No. Well some small things might add up, but I think you don't see that untill the TB range. >For example I have 1x 40G, 1x80G, 1x120G XFS partitions on the same >fileserver with 384M RAM. Is this enough? If people are reading _and_ writing this data at random over all the disks then it is more then enough. If there is more data read then written, buying more ram is a sane thing to do. If this is for a internet server behind a 64Kbit line that won't help the people leeching from it :-) A politically correct answer would be "it depends". Cheers -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Wed Dec 5 02:08:34 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5A8YR26181 for linux-xfs-outgoing; Wed, 5 Dec 2001 02:08:34 -0800 Received: from red.csi.cam.ac.uk (exim@red.csi.cam.ac.uk [131.111.8.70]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5A8Po26158 for ; Wed, 5 Dec 2001 02:08:25 -0800 Received: from tmd.christs.cam.ac.uk ([131.111.219.75] helo=tmd.cam.ac.uk) by red.csi.cam.ac.uk with esmtp (Exim 3.22 #1) id 16BY2C-0005tz-00; Wed, 05 Dec 2001 09:08:08 +0000 Message-Id: <5.1.0.14.2.20011205090142.04ab5b20@pop.cus.cam.ac.uk> X-Sender: aia21@pop.cus.cam.ac.uk X-Mailer: QUALCOMM Windows Eudora Version 5.1 Date: Wed, 05 Dec 2001 09:08:12 +0000 To: Nathan Scott From: Anton Altaparmakov Subject: Re: [PATCH] Revised extended attributes interface Cc: Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com In-Reply-To: <20011205143209.C44610@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 03:32 05/12/01, Nathan Scott wrote: >Here is the revised interface. I believe it takes into account >the issues raised so far - further suggestions are also welcome, >of course. Hi, Looks good to me. Just one tiny point: you seem to like setting error=xyz; a lot which is completely unnecessary some times. Any particular reason? Here is an example of what I mean: >+static long >+setxattr(struct dentry *d, char *name, void *value, size_t size, int flags) >+{ >+ int error; >+ void *kvalue; >+ char kname[XATTR_NAME_MAX + 1]; >+ >+ error = -EINVAL; >+ if (flags & ~(XATTR_CREATE|XATTR_REPLACE)) >+ return error; Why not: + if (flags & ~(XATTR_CREATE|XATTR_REPLACE)) + return -EINVAL; >+ >+ error = -EFAULT; >+ if (copy_from_user(kname, name, XATTR_NAME_MAX)) >+ return error; + if (copy_from_user(kname, name, XATTR_NAME_MAX)) + return -EFAULT; >+ kname[XATTR_NAME_MAX] = '\0'; >+ >+ kvalue = xattr_alloc(size, XATTR_SIZE_MAX); >+ if (IS_ERR(kvalue)) >+ return PTR_ERR(kvalue); >+ >+ error = -EFAULT; >+ if (size > 0 && copy_from_user(kvalue, value, size)) { >+ xattr_free(kvalue, size); >+ return error; >+ } + if (size > 0 && copy_from_user(kvalue, value, size)) { + xattr_free(kvalue, size); + return -EFAULT; + } Shorter, faster in the common non-error path, and looks nicer, although the latter is probably a matter of personal preference. (-; Best regards, Anton -- "I've not lost my mind. It's backed up on tape somewhere." - Unknown -- Anton Altaparmakov (replace at with @) Linux NTFS Maintainer / WWW: http://linux-ntfs.sf.net/ ICQ: 8561279 / WWW: http://www-stu.christs.cam.ac.uk/~aia21/ From owner-linux-xfs@oss.sgi.com Wed Dec 5 03:56:56 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5BuuW29195 for linux-xfs-outgoing; Wed, 5 Dec 2001 03:56:56 -0800 Received: from anime.net (root@anime.net [63.172.78.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Buro29173 for ; Wed, 5 Dec 2001 03:56:53 -0800 Received: from localhost (goemon@localhost) by anime.net (8.11.6/8.9.3) with ESMTP id fB5Auq428882 for ; Wed, 5 Dec 2001 02:56:52 -0800 Date: Wed, 5 Dec 2001 02:56:52 -0800 (PST) From: Dan Hollis To: linux xfs ml Subject: How to fsck read-only-mounted root filesystem? Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk How can one xfs_check/xfs_repair a root filesystem? If you try this, even from singleuser mode where / is mounted read-only, the utilities whinge about the filesystem being mounted, and refuse to touch the disk at all. I have an xfs filesystem with some corruption from (I think) bad dma. There's some directories which xfs won't let me remove. And with no way to repair the filesystem :-( :-( this sux. Seems to me that repairing a read-only mounted filesystem should be allowed?!? e2fsck allows it. -Dan -- [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Wed Dec 5 05:39:48 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5Ddm801167 for linux-xfs-outgoing; Wed, 5 Dec 2001 05:39:48 -0800 Received: from camelot.virtualavalon.net (127bus50.tampabay.rr.com [24.94.127.50]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Ddho01145 for ; Wed, 5 Dec 2001 05:39:44 -0800 Received: from arthur.virtualavalon.net (arthur.virtualavalon.net [172.20.1.10]) by camelot.virtualavalon.net (8.12.0/8.12.0) with ESMTP id fB5Cdfrh023992 for ; Wed, 5 Dec 2001 07:39:41 -0500 Received: from tampabay.rr.com (wayfarer.virtualavalon.net [172.20.1.15]) by arthur.virtualavalon.net (8.12.0/8.12.0) with ESMTP id fB5CdYL5001359 for ; Wed, 5 Dec 2001 07:39:36 -0500 (EST) Message-ID: <3C0E157D.50502@tampabay.rr.com> Date: Wed, 05 Dec 2001 07:39:25 -0500 From: "Jesse W. Asher" User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:0.9.4) Gecko/20011019 Netscape6/6.2 X-Accept-Language: en-us MIME-Version: 1.0 To: linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? References: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk You have to remount it on the fly. Look at the man page for "mount" and look at the remount option. You should be able to do something like "mount -o remount,rw /" or some such. I'm not sure this is the exact syntax, but you should be able to figure it out..... Dan Hollis wrote: >How can one xfs_check/xfs_repair a root filesystem? > >If you try this, even from singleuser mode where / is mounted read-only, >the utilities whinge about the filesystem being mounted, and refuse to >touch the disk at all. > >I have an xfs filesystem with some corruption from (I think) bad dma. >There's some directories which xfs won't let me remove. And with no way >to repair the filesystem :-( :-( this sux. > >Seems to me that repairing a read-only mounted filesystem should be >allowed?!? e2fsck allows it. > >-Dan > -- Jesse W. Asher "They that can give up essential liberty to purchase a little temporary safety, deserve neither liberty or safety." - Benjamin Franklin From owner-linux-xfs@oss.sgi.com Wed Dec 5 05:58:43 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5DwhX01665 for linux-xfs-outgoing; Wed, 5 Dec 2001 05:58:43 -0800 Received: from ADSL-Bergs.RZ.RWTH-Aachen.DE (adsl-bergs.rz.RWTH-Aachen.DE [137.226.80.218]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Dwao01643 for ; Wed, 5 Dec 2001 05:58:37 -0800 Received: from [192.168.1.2] (helo=ralf) by ADSL-Bergs.RZ.RWTH-Aachen.DE with smtp (Exim 3.12 #1) id 16Bbcl-0002S3-00; Wed, 05 Dec 2001 13:58:07 +0100 From: "Ralf G. R. Bergs" To: "linux xfs ml" Cc: "Jesse W. Asher" Date: Wed, 05 Dec 2001 13:58:06 +0100 Reply-To: "Ralf G. R. Bergs" X-Mailer: PMMail 2000 Professional (2.20.2370) For Windows 2000 (5.0.2195;2) In-Reply-To: <3C0E157D.50502@tampabay.rr.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Subject: Re: How to fsck read-only-mounted root filesystem? Message-Id: Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk [Quote repaired] On Wed, 05 Dec 2001 07:39:25 -0500, Jesse W. Asher wrote: >Dan Hollis wrote: > >>How can one xfs_check/xfs_repair a root filesystem? >> >>If you try this, even from singleuser mode where / is mounted read-only, >>the utilities whinge about the filesystem being mounted, and refuse to >>touch the disk at all. >> >>I have an xfs filesystem with some corruption from (I think) bad dma. >>There's some directories which xfs won't let me remove. And with no way >>to repair the filesystem :-( :-( this sux. >> >>Seems to me that repairing a read-only mounted filesystem should be >>allowed?!? e2fsck allows it. > >You have to remount it on the fly. Look at the man page for "mount" and >look at the remount option. You should be able to do something like >"mount -o remount,rw /" or some such. I'm not sure this is the exact >syntax, but you should be able to figure it out..... I guess you should have another look at the original message -- your answer simply does not apply to Dan's question. -- Verkaufe Original-BMW-Raeder: L I N U X .~. http://adsl-bergs.rz.rwth-aachen.de/~rabe The Choice /V\ of a GNU /( )\ Generation ^^-^^ From owner-linux-xfs@oss.sgi.com Wed Dec 5 06:06:02 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5E62K02036 for linux-xfs-outgoing; Wed, 5 Dec 2001 06:06:02 -0800 Received: from relay.xlink.net (relay.xlink.net [193.141.40.4]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5E5vo02014 for ; Wed, 5 Dec 2001 06:05:57 -0800 Received: from lizard.webland.de (lizard.webland.de [194.122.76.201]) by relay.xlink.net (8.9.3/8.8.7) with ESMTP id OAA06051; Wed, 5 Dec 2001 14:05:24 +0100 (MET) Received: (from uucp@localhost) by lizard.webland.de (8.8.8/8.8.7) id OAA01186; Wed, 5 Dec 2001 14:05:18 +0100 (MET) >Received: from mobile.sauter-bc.com (unknown [10.1.6.21]) by basel1.sauter-bc.com (Postfix) with ESMTP id 5421F57306; Wed, 5 Dec 2001 14:04:49 +0100 (CET) Received: from ch.sauter-bc.com (support.cad.sba [10.1.200.117]) by mobile.sauter-bc.com (Postfix) with ESMTP id B994125835; Wed, 5 Dec 2001 14:04:44 +0100 (CET) Message-ID: <3C0E1B6C.F4518BDD@ch.sauter-bc.com> Date: Wed, 05 Dec 2001 14:04:44 +0100 From: Simon Matter Organization: Sauter AG, Basel X-Mailer: Mozilla 4.77 [de] (X11; U; Linux 2.2.19-6.2.12 i686) X-Accept-Language: de-CH, en MIME-Version: 1.0 To: Dan Hollis Cc: linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? References: Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Dan Hollis schrieb: > > How can one xfs_check/xfs_repair a root filesystem? Boot from a bootable CD and perform the repair on the unmounted filesystem. Have a look at this http://lbt.linuxcare.com/index.epl HTH. -Simon > > If you try this, even from singleuser mode where / is mounted read-only, > the utilities whinge about the filesystem being mounted, and refuse to > touch the disk at all. > > I have an xfs filesystem with some corruption from (I think) bad dma. > There's some directories which xfs won't let me remove. And with no way > to repair the filesystem :-( :-( this sux. > > Seems to me that repairing a read-only mounted filesystem should be > allowed?!? e2fsck allows it. > > -Dan > -- > [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Wed Dec 5 07:03:33 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5F3XT03761 for linux-xfs-outgoing; Wed, 5 Dec 2001 07:03:33 -0800 Received: from minnie.omroep.nl (minnie.omroep.nl [145.58.30.4]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5F3So03737 for ; Wed, 5 Dec 2001 07:03:28 -0800 Received: (from smap@localhost) by minnie.omroep.nl (SGI-8.9.3/8.9.3) id PAA10102 for ; Wed, 5 Dec 2001 15:03:21 +0100 (CET) Received: from kwek.omroep.nl (145.58.31.3 "EHLO kwek.omroep.nl") by minnie.omroep.nl with ESMTP (smap v3.0 nederlandse publieke omroep) id xma4822510; Wed, 5 Dec 01 15:03:19 +0100 Received: from localhost (matthijs@localhost) by kwek.omroep.nl (SGI-8.9.3/8.9.3) with ESMTP id PAA54741 for ; Wed, 5 Dec 2001 15:03:16 +0100 (CET) X-Authentication-Warning: kwek.omroep.nl: matthijs owned process doing -bs Date: Wed, 5 Dec 2001 15:03:16 +0100 From: Matthijs van der Klip X-X-Sender: To: Linux XFS Mailing List Subject: using xfsdump to synchronise filesystems? Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, I know this is a bit of a crazy idea, but I'd like to know if something like this is (remotely) possible... 1) Take a server with a 10Gb (max) volume. This volume is live: files are being added/deleted through ftp (Read/Write). 2) Take x clients which need to receive a synchronised copy of the mentioned volume above. Contents have to be served out through http (Read only). Would it be possible to say make a level 0 dump every night, a level 1 dump every hour, and a level 2 dump every 5 minutes (or use even more levels?) and then restore these dumps on the clients to keep their volumes synchronised? Reason I'm asking this is I'm looking for a way to distribute only changes in a filesystem. Rsync has to scan the whole filesystem and just takes to long. Groet, -- Matthijs van der Klip - Unix Administrator NOS - Dutch Public Broadcasting Organisation http://www.omroep.nl From owner-linux-xfs@oss.sgi.com Wed Dec 5 08:00:49 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5G0nL09281 for linux-xfs-outgoing; Wed, 5 Dec 2001 08:00:49 -0800 Received: from real-uma.mit.edu (REAL-UMA.MIT.EDU [18.82.0.129]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5G0jo09256 for ; Wed, 5 Dec 2001 08:00:45 -0800 Received: (qmail 3444 invoked by uid 1001); 5 Dec 2001 15:00:43 -0000 Date: Wed, 5 Dec 2001 10:00:43 -0500 From: Xianglong Yuan To: linux-xfs@oss.sgi.com Subject: Files on XFS not safe?! Message-ID: <20011205100043.A3437@real-uma.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.3.20i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, there, I've just transferred my old ext2 FS into XFS FS and did the following test. What I got surprises me. First, I modify a file with vim, save it, and go back to bash shell. Then, I hit the power button to crash the system instantly. Now I restart the system and the XFS FS comes back gracefully. However, when I try to open the just saved file, nothing is there, not the new modified content, not even the old content. All the content in the file is wiped off by a stream of weird symbols. Is the XFS FS supposed to behavior like that? It seems to me XFS is really not a safe FS to count on when system crashing. I presumed that I should have my old content back. Did I asked too much here? Thanks in advance. Xianglong Yuan From owner-linux-xfs@oss.sgi.com Wed Dec 5 08:07:19 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5G7Jo09638 for linux-xfs-outgoing; Wed, 5 Dec 2001 08:07:19 -0800 Received: from chaos.egr.duke.edu (chaos.egr.duke.edu [152.3.195.82]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5G7Go09615 for ; Wed, 5 Dec 2001 08:07:16 -0800 Received: from localhost (jlb@localhost) by chaos.egr.duke.edu (8.11.6/8.11.6) with ESMTP id fB5F7E717960; Wed, 5 Dec 2001 10:07:14 -0500 X-Authentication-Warning: chaos.egr.duke.edu: jlb owned process doing -bs Date: Wed, 5 Dec 2001 10:07:14 -0500 (EST) From: Joshua Baker-LePain X-X-Sender: To: Xianglong Yuan cc: Subject: Re: Files on XFS not safe?! In-Reply-To: <20011205100043.A3437@real-uma.mit.edu> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 5 Dec 2001 at 10:00am, Xianglong Yuan wrote > First, I modify a file with vim, save it, and go back to bash > shell. Then, I hit the power button to crash the system > instantly. Now I restart the system and the XFS FS comes back > gracefully. However, when I try to open the just saved file, > nothing is there, not the new modified content, not even the old > content. All the content in the file is wiped off by a stream > of weird symbols. Is the XFS FS supposed to behavior like that? > It seems to me XFS is really not a safe FS to count on when > system crashing. I presumed that I should have my old content > back. Did I asked too much here? Thanks in advance. > http://oss.sgi.com/projects/xfs/faq.html#nulls -- Joshua Baker-LePain Department of Biomedical Engineering Duke University From owner-linux-xfs@oss.sgi.com Wed Dec 5 08:09:23 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5G9NC09801 for linux-xfs-outgoing; Wed, 5 Dec 2001 08:09:23 -0800 Received: from mail.dkp.com ([204.191.16.3]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5G9Ko09778 for ; Wed, 5 Dec 2001 08:09:20 -0800 Received: from ranma.dkp.com (ranma.dkp.com [205.150.40.12]) by mail.dkp.com (Postfix) with ESMTP id 4AE761AB36 for ; Wed, 5 Dec 2001 10:09:19 -0500 (EST) Received: by ranma.dkp.com (Postfix, from userid 168) id CAC1B5AA0B; Wed, 5 Dec 2001 10:09:18 -0500 (EST) Date: Wed, 5 Dec 2001 10:09:18 -0500 From: Andrew Klaassen To: linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! Message-ID: <20011205100918.B3832@dkp.com> Mail-Followup-To: linux-xfs@oss.sgi.com References: <20011205100043.A3437@real-uma.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20011205100043.A3437@real-uma.mit.edu> User-Agent: Mutt/1.3.23i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 10:00:43AM -0500, Xianglong Yuan wrote: > First, I modify a file with vim, save it, and go back to bash > shell. Then, I hit the power button to crash the system > instantly. Now I restart the system and the XFS FS comes back > gracefully. However, when I try to open the just saved file, > nothing is there, not the new modified content, not even the > old content. All the content in the file is wiped off by a > stream of weird symbols. Is the XFS FS supposed to behavior > like that? It seems to me XFS is really not a safe FS to count > on when system crashing. I presumed that I should have my old > content back. Did I asked too much here? Thanks in advance. http://oss.sgi.com/projects/xfs/faq.html#nulls Andrew Klaassen From owner-linux-xfs@oss.sgi.com Wed Dec 5 08:47:48 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5Glmg12696 for linux-xfs-outgoing; Wed, 5 Dec 2001 08:47:48 -0800 Received: from real-uma.mit.edu (REAL-UMA.MIT.EDU [18.82.0.129]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Glgo12670 for ; Wed, 5 Dec 2001 08:47:43 -0800 Received: (qmail 3450 invoked by uid 1001); 5 Dec 2001 15:47:41 -0000 Date: Wed, 5 Dec 2001 10:47:41 -0500 From: Xianglong Yuan To: Paulo Sergio Lemes Queiroz Cc: linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! Message-ID: <20011205104741.A3446@real-uma.mit.edu> References: <20011205100043.A3437@real-uma.mit.edu> <3C0E3A22.C504B179@definitylinux.com.br> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3C0E3A22.C504B179@definitylinux.com.br> User-Agent: Mutt/1.3.20i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >-Paulo Sergio Lemes Queiroz [05 Dec 2001 13:15 -0200] wrote: > > http://oss.sgi.com/projects/xfs/faq.html#nulls Thanks. Is this true for all the journaling FS, say ReiserFS, JFS? Xianglong Yuan > > > Xianglong Yuan wrote: > > > Hi, there, > > > > I've just transferred my old ext2 FS into XFS FS and did the > > following test. What I got surprises me. > > > > First, I modify a file with vim, save it, and go back to bash > > shell. Then, I hit the power button to crash the system > > instantly. Now I restart the system and the XFS FS comes back > > gracefully. However, when I try to open the just saved file, > > nothing is there, not the new modified content, not even the old > > content. All the content in the file is wiped off by a stream > > of weird symbols. Is the XFS FS supposed to behavior like that? > > It seems to me XFS is really not a safe FS to count on when > > system crashing. I presumed that I should have my old content > > back. Did I asked too much here? Thanks in advance. > > > > Xianglong Yuan > > -- > Paulo Sergio Lemes Queiroz > > Analista de Sistemas > Definity Linux | Esc Telecomunica??es > > From owner-linux-xfs@oss.sgi.com Wed Dec 5 08:57:05 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5Gv5i13007 for linux-xfs-outgoing; Wed, 5 Dec 2001 08:57:05 -0800 Received: from porsta.cs.Helsinki.FI (root@porsta.cs.Helsinki.FI [128.214.48.124]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Gv0o12983 for ; Wed, 5 Dec 2001 08:57:01 -0800 Received: from melkki.cs.Helsinki.FI (sslwrap@localhost [127.0.0.1]) by porsta.cs.Helsinki.FI (8.11.6/8.11.6) with ESMTP id fB5FurE02369; Wed, 5 Dec 2001 17:56:54 +0200 Received: (from hhaataja@localhost) by melkki.cs.Helsinki.FI (8.11.6/8.11.2) id fB5FulQ05984; Wed, 5 Dec 2001 17:56:47 +0200 Date: Wed, 5 Dec 2001 17:56:47 +0200 From: Harri Haataja To: Xianglong Yuan Cc: Paulo Sergio Lemes Queiroz , linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! Message-ID: <20011205175647.A4895@cs.helsinki.fi> Mail-Followup-To: Xianglong Yuan , Paulo Sergio Lemes Queiroz , linux-xfs@oss.sgi.com References: <20011205100043.A3437@real-uma.mit.edu> <3C0E3A22.C504B179@definitylinux.com.br> <20011205104741.A3446@real-uma.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20011205104741.A3446@real-uma.mit.edu>; from yuanx@mit.edu on Wed, Dec 05, 2001 at 10:47:41AM -0500 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 10:47:41AM -0500, Xianglong Yuan wrote: > >-Paulo Sergio Lemes Queiroz [05 Dec 2001 13:15 -0200] wrote: > > > > http://oss.sgi.com/projects/xfs/faq.html#nulls > > Thanks. Is this true for all the journaling FS, say ReiserFS, > JFS? Maybe I might bring up this http://kerneltrap.com/article.php?sid=389 -- Real Programmers don't write in FORTRAN. FORTRAN is for wimp engineers who wear white socks. They get excited over finite state analysis and nuclear reactor simulation. From owner-linux-xfs@oss.sgi.com Wed Dec 5 09:14:28 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5HES713554 for linux-xfs-outgoing; Wed, 5 Dec 2001 09:14:28 -0800 Received: from real-uma.mit.edu (REAL-UMA.MIT.EDU [18.82.0.129]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5HENo13532 for ; Wed, 5 Dec 2001 09:14:23 -0800 Received: (qmail 3456 invoked by uid 1001); 5 Dec 2001 16:14:21 -0000 Date: Wed, 5 Dec 2001 11:14:21 -0500 From: Xianglong Yuan To: Joshua Baker-LePain Cc: linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! Message-ID: <20011205111421.B3446@real-uma.mit.edu> References: <20011205100043.A3437@real-uma.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.3.20i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >-Joshua Baker-LePain [05 Dec 2001 10:07 -0500] wrote: > > On Wed, 5 Dec 2001 at 10:00am, Xianglong Yuan wrote > > > First, I modify a file with vim, save it, and go back to bash > > shell. Then, I hit the power button to crash the system > > instantly. Now I restart the system and the XFS FS comes back > > gracefully. However, when I try to open the just saved file, > > nothing is there, not the new modified content, not even the old > > content. All the content in the file is wiped off by a stream > > of weird symbols. Is the XFS FS supposed to behavior like that? > > It seems to me XFS is really not a safe FS to count on when > > system crashing. I presumed that I should have my old content > > back. Did I asked too much here? Thanks in advance. > > > http://oss.sgi.com/projects/xfs/faq.html#nulls Is there any way to recover the old content? It is still at somewhere in the disk, isn't it? > > -- > Joshua Baker-LePain > Department of Biomedical Engineering > Duke University > Xianglong Yuan -- Dept. Materials Science and Engineering MIT, Room 13-4050 Cambridge, MA 02139 From owner-linux-xfs@oss.sgi.com Wed Dec 5 09:25:33 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5HPXl16411 for linux-xfs-outgoing; Wed, 5 Dec 2001 09:25:33 -0800 Received: from mclean.mail.mindspring.net (mclean.mail.mindspring.net [207.69.200.57]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5HPPo16388 for ; Wed, 5 Dec 2001 09:25:25 -0800 Received: from user-uini6c4.dsl.mindspring.com ([165.121.25.132] helo=waltsathlon.localhost.net) by mclean.mail.mindspring.net with esmtp (Exim 3.33 #1) id 16BerM-0002ti-00 for linux-xfs@oss.sgi.com; Wed, 05 Dec 2001 11:25:24 -0500 Received: from mindspring.com (localhost.localdomain [127.0.0.1]) by waltsathlon.localhost.net (Postfix) with ESMTP id 56AC7D066334; Wed, 5 Dec 2001 08:15:56 -0800 (PST) Message-ID: <3C0E483C.9020406@mindspring.com> Date: Wed, 05 Dec 2001 08:15:56 -0800 From: Walt H User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.6+) Gecko/20011202 X-Accept-Language: en-us MIME-Version: 1.0 To: Xianglong Yuan Cc: Paulo Sergio Lemes Queiroz , linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! References: <20011205100043.A3437@real-uma.mit.edu> <3C0E3A22.C504B179@definitylinux.com.br> <20011205104741.A3446@real-uma.mit.edu> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk The potential is certainly there. Most jfs use delayed writing. When you save something to disk, it isn't immediately flushed to disk. Had you ran 'sync' prior to pushing the power button, perhaps the file would've been intact? If you use EXT3 and set it to full data journaling mode, it might've made it that way also. In that mode, EXT3 journals meta data and file data. -Walt Xianglong Yuan wrote: >>-Paulo Sergio Lemes Queiroz [05 Dec 2001 13:15 -0200] wrote: >> >>http://oss.sgi.com/projects/xfs/faq.html#nulls >> > > Thanks. Is this true for all the journaling FS, say ReiserFS, > JFS? > > Xianglong Yuan > > >> >>Xianglong Yuan wrote: >> >> >>>Hi, there, >>> >>>I've just transferred my old ext2 FS into XFS FS and did the >>>following test. What I got surprises me. >>> >>>First, I modify a file with vim, save it, and go back to bash >>>shell. Then, I hit the power button to crash the system >>>instantly. Now I restart the system and the XFS FS comes back >>>gracefully. However, when I try to open the just saved file, >>>nothing is there, not the new modified content, not even the old >>>content. All the content in the file is wiped off by a stream >>>of weird symbols. Is the XFS FS supposed to behavior like that? >>>It seems to me XFS is really not a safe FS to count on when >>>system crashing. I presumed that I should have my old content >>>back. Did I asked too much here? Thanks in advance. >>> >>>Xianglong Yuan >>> >>-- >>Paulo Sergio Lemes Queiroz >> >>Analista de Sistemas >>Definity Linux | Esc Telecomunica??es >> >> >> > From owner-linux-xfs@oss.sgi.com Wed Dec 5 09:30:21 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5HULj16626 for linux-xfs-outgoing; Wed, 5 Dec 2001 09:30:21 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5HUEo16601 for ; Wed, 5 Dec 2001 09:30:14 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id IAA24563 for ; Wed, 5 Dec 2001 08:30:01 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id KAA3770608; Wed, 5 Dec 2001 10:28:54 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id KAA74042; Wed, 5 Dec 2001 10:28:54 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5GSXF24181; Wed, 5 Dec 2001 10:28:33 -0600 Subject: Re: Files on XFS not safe?! From: Steve Lord To: Xianglong Yuan Cc: Joshua Baker-LePain , linux-xfs@oss.sgi.com In-Reply-To: <20011205111421.B3446@real-uma.mit.edu> References: <20011205100043.A3437@real-uma.mit.edu> <20011205111421.B3446@real-uma.mit.edu> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 10:28:33 -0600 Message-Id: <1007569713.22718.10.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 10:14, Xianglong Yuan wrote: > >-Joshua Baker-LePain [05 Dec 2001 10:07 -0500] wrote: > > > > On Wed, 5 Dec 2001 at 10:00am, Xianglong Yuan wrote > > > > > First, I modify a file with vim, save it, and go back to bash > > > shell. Then, I hit the power button to crash the system > > > instantly. Now I restart the system and the XFS FS comes back > > > gracefully. However, when I try to open the just saved file, > > > nothing is there, not the new modified content, not even the old > > > content. All the content in the file is wiped off by a stream > > > of weird symbols. Is the XFS FS supposed to behavior like that? > > > It seems to me XFS is really not a safe FS to count on when > > > system crashing. I presumed that I should have my old content > > > back. Did I asked too much here? Thanks in advance. > > > > > http://oss.sgi.com/projects/xfs/faq.html#nulls > > Is there any way to recover the old content? It is still at > somewhere in the disk, isn't it? No, it was in the memory on your machine when you turned off the power, now it is gone. The problem here is partially the editor, partially you, and partially XFS. o The editor because it does not use a call such as fsync on the new file before renaming it over the original. o You for pulling the power at the wrong moment (OK, it could be the power company, or the fuse box too). o XFS for allowing a size update out to disk before the file data, not that this helps too much, you would just get a zero length file instead of a file with NULLs in it. Steve > > > > > -- > > Joshua Baker-LePain > > Department of Biomedical Engineering > > Duke University > > > > Xianglong Yuan > -- > Dept. Materials Science and Engineering > MIT, Room 13-4050 > Cambridge, MA 02139 -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 09:31:05 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5HV5p16765 for linux-xfs-outgoing; Wed, 5 Dec 2001 09:31:05 -0800 Received: from eclectic.kluge.net (IDENT:root@dsl092-071-242.bos1.dsl.speakeasy.net [66.92.71.242]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5HV2o16742 for ; Wed, 5 Dec 2001 09:31:02 -0800 Received: (from felicity@localhost) by eclectic.kluge.net (8.11.6/8.11.6) id fB5GUwn26237; Wed, 5 Dec 2001 11:30:58 -0500 Date: Wed, 5 Dec 2001 11:30:58 -0500 From: Theo Van Dinter To: Russell Cattelan Cc: linux-xfs@oss.sgi.com Subject: Re: html filter test Message-ID: <20011205113057.A26046@kluge.net> References: <1007510005.2534.2.camel@scare> <20011205000427.GG1152@gwyn.tux.org> <1007511537.2534.11.camel@scare> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <1007511537.2534.11.camel@scare>; from cattelan@thebarn.com on Tue, Dec 04, 2001 at 06:18:57PM -0600 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Tue, Dec 04, 2001 at 06:18:57PM -0600, Russell Cattelan wrote: > script all typed up > > http://www.kluge.net/mailfiltering/strip_html.txt Wow, I'm happy to see people using this! :) While this remains off-topic for the XFS list (sorry), you may want to poke at http://www.kluge.net/mailfiltering/ in general. It has a few other fun procmail things in there. The top section is specific to my server (accessdb, sendmail header blocking, etc,) but the bottom parts are generic. This reminds me, I have a few new things to put in there, like spamdunk (http://www3.sympatico.ca/walter.dnes/email/spamdunk/index.html). :) -- Randomly Generated Tagline: "From what I'd seen of British TV, some shows use the word bastard like I use a comma." - J. Michael Straczynski From owner-linux-xfs@oss.sgi.com Wed Dec 5 09:37:18 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5HbIj16934 for linux-xfs-outgoing; Wed, 5 Dec 2001 09:37:18 -0800 Received: from mustard.heime.net (mustard.heime.net [194.234.65.222]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5HbFo16912 for ; Wed, 5 Dec 2001 09:37:15 -0800 Received: from localhost (roy@localhost) by mustard.heime.net (8.11.6/8.9.3) with ESMTP id fB5Gatw02484 for ; Wed, 5 Dec 2001 17:36:55 +0100 Date: Wed, 5 Dec 2001 17:36:55 +0100 (CET) From: Roy Sigurd Karlsbakk X-Sender: To: XFS Mailing list Subject: XFS readahead setting??? Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi Does XFS have any readahead setting on its own? I've tried the /proc/sys/vm/(max|min)-readahead setting in Linux without any success. I need somewhere between 4 and 16 MB readahead per file to be able to read a large number of large files concurrently, and to reduce seeks to a minimum. thanks roy -- Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA Computers are like air conditioners. They stop working when you open Windows. From owner-linux-xfs@oss.sgi.com Wed Dec 5 10:00:23 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5I0ND17373 for linux-xfs-outgoing; Wed, 5 Dec 2001 10:00:23 -0800 Received: from camelot.virtualavalon.net (127bus50.tampabay.rr.com [24.94.127.50]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5I0Do17345 for ; Wed, 5 Dec 2001 10:00:13 -0800 Received: from arthur.virtualavalon.net (arthur.virtualavalon.net [172.20.1.10]) by camelot.virtualavalon.net (8.12.0/8.12.0) with ESMTP id fB5H0Brh024091 for ; Wed, 5 Dec 2001 12:00:11 -0500 Received: from tampabay.rr.com (wayfarer.virtualavalon.net [172.20.1.15]) by arthur.virtualavalon.net (8.12.0/8.12.0) with ESMTP id fB5H06L5001779 for ; Wed, 5 Dec 2001 12:00:06 -0500 (EST) Message-ID: <3C0E528D.4040703@tampabay.rr.com> Date: Wed, 05 Dec 2001 11:59:57 -0500 From: "Jesse W. Asher" User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:0.9.4) Gecko/20011019 Netscape6/6.2 X-Accept-Language: en-us MIME-Version: 1.0 To: linux-xfs@oss.sgi.com Subject: Re: How to fsck read-only-mounted root filesystem? References: Content-Disposition: inline Content-Type: text/plain Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Sorry 'bout that! I misunderstood and thought it was complaining because it was read only. In this case there are a couple of alternatives. The easiest would be to boot off of a CD or floppy. Probably the CD would be best because the floppy is bound to use the root filesystem off the hard drive and you might still run into the same problem. Another alternative is to do a network boot, but if you don't already have that set up it will be a long and exhausting process to get it set up just to repair the filesystem.. In short, there doesn't seem to be a painless way to do this. The CD isn't likley to have XFS support built in. If you built the XFS stuff as modules, you could possibly boot off the CD and then load the XFS modules using insmod. Also, if you haven't tried the remount option I originally mentioned, you may want to try it anyway even though the utilities complain about touching a mounted filesystem. I don't always trust error messages to be accurate and there is a possibility that it doesn't like the fact that it is read-only rather than just being mounted. You won't know for sure until you try and that is an easy thing to try. Just a thought. I seem to remember there being instructions on how to build an XFS-based root filesystem and there are likely to be clues for fixing your current situation in that documentation. May be worthwhile checking out to see what you can gleen from various documents.... Dan Hollis wrote: >i dont think you read my post very closely... ;P > >xfs_check/xfs_repair refuse to touch any FS if it's mounted, even if it's >mounted read-only... > >-Dan > >On Wed, 5 Dec 2001, Jesse W. Asher wrote: > >>You have to remount it on the fly. Look at the man page for "mount" and >>look at the remount option. You should be able to do something like >>"mount -o remount,rw /" or some such. I'm not sure this is the exact >>syntax, but you should be able to figure it out..... >> >>Dan Hollis wrote: >> >>>How can one xfs_check/xfs_repair a root filesystem? >>> >>>If you try this, even from singleuser mode where / is mounted read-only, >>>the utilities whinge about the filesystem being mounted, and refuse to >>>touch the disk at all. >>> >>>I have an xfs filesystem with some corruption from (I think) bad dma. >>>There's some directories which xfs won't let me remove. And with no way >>>to repair the filesystem :-( :-( this sux. >>> >>>Seems to me that repairing a read-only mounted filesystem should be >>>allowed?!? e2fsck allows it. >>> >>>-Dan >>> >> > -- Jesse W. Asher "They that can give up essential liberty to purchase a little temporary safety, deserve neither liberty or safety." - Benjamin Franklin [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Wed Dec 5 10:03:35 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5I3Zt17572 for linux-xfs-outgoing; Wed, 5 Dec 2001 10:03:35 -0800 Received: from real-uma.mit.edu (REAL-UMA.MIT.EDU [18.82.0.129]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5I3Ro17550 for ; Wed, 5 Dec 2001 10:03:27 -0800 Received: (qmail 3463 invoked by uid 1001); 5 Dec 2001 17:03:25 -0000 Date: Wed, 5 Dec 2001 12:03:25 -0500 From: Xianglong Yuan To: Steve Lord Cc: linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! Message-ID: <20011205120325.C3446@real-uma.mit.edu> References: <20011205100043.A3437@real-uma.mit.edu> <20011205111421.B3446@real-uma.mit.edu> <1007569713.22718.10.camel@jen.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1007569713.22718.10.camel@jen.americas.sgi.com> User-Agent: Mutt/1.3.20i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >-Steve Lord [05 Dec 2001 10:28 -0600] wrote: > > On Wed, 2001-12-05 at 10:14, Xianglong Yuan wrote: > > >-Joshua Baker-LePain [05 Dec 2001 10:07 -0500] wrote: > > > > > > On Wed, 5 Dec 2001 at 10:00am, Xianglong Yuan wrote > > > > > > > First, I modify a file with vim, save it, and go back to bash > > > > shell. Then, I hit the power button to crash the system > > > > instantly. Now I restart the system and the XFS FS comes back > > > > gracefully. However, when I try to open the just saved file, > > > > nothing is there, not the new modified content, not even the old > > > > content. All the content in the file is wiped off by a stream > > > > of weird symbols. Is the XFS FS supposed to behavior like that? > > > > It seems to me XFS is really not a safe FS to count on when > > > > system crashing. I presumed that I should have my old content > > > > back. Did I asked too much here? Thanks in advance. > > > > > > > http://oss.sgi.com/projects/xfs/faq.html#nulls > > > > Is there any way to recover the old content? It is still at > > somewhere in the disk, isn't it? > > No, it was in the memory on your machine when you turned off the power, > now it is gone. The problem here is partially the editor, partially you, > and partially XFS. > > o The editor because it does not use a call such as fsync on the new > file before renaming it over the original. > > o You for pulling the power at the wrong moment (OK, it could be the > power company, or the fuse box too). > > o XFS for allowing a size update out to disk before the file data, not > that this helps too much, you would just get a zero length file > instead of a file with NULLs in it. I presume there is a log that has the information when the data-flashing is scheduled to start, whether it has done or not, and what the original file state (meta-data?) is. Now, if the system crashed before the data-flashing is done, could it be possible to trace the log back to the point before the unfinished data-flashing and give the original file state (meta-data?) back instead of the meta-data of the new file as it gives nothing? (I believe the content of the old file is still on the disk and not zeroed by other program, and could be recovered if we know its locations (meta-data?).) Update the new meta-data only after a successful data-flashing. Am I missing something here? Xianglong > > Steve > > > > > > > > > -- > > > Joshua Baker-LePain > > > Department of Biomedical Engineering > > > Duke University > > > > > > > Xianglong Yuan > > -- > > Dept. Materials Science and Engineering > > MIT, Room 13-4050 > > Cambridge, MA 02139 > -- > > Steve Lord voice: +1-651-683-3511 > Principal Engineer, Filesystem Software email: lord@sgi.com -- Dept. Materials Science and Engineering MIT, Room 13-4050 Cambridge, MA 02139 From owner-linux-xfs@oss.sgi.com Wed Dec 5 10:16:57 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5IGv318734 for linux-xfs-outgoing; Wed, 5 Dec 2001 10:16:57 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5IGpo18712 for ; Wed, 5 Dec 2001 10:16:51 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id JAA08699 for ; Wed, 5 Dec 2001 09:16:52 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3747524; Wed, 5 Dec 2001 11:15:31 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id LAA48825; Wed, 5 Dec 2001 11:15:31 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5HFAH24257; Wed, 5 Dec 2001 11:15:10 -0600 Subject: Re: XFS readahead setting??? From: Steve Lord To: Roy Sigurd Karlsbakk Cc: XFS Mailing list In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 11:15:10 -0600 Message-Id: <1007572510.22679.14.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 10:36, Roy Sigurd Karlsbakk wrote: > hi > > Does XFS have any readahead setting on its own? I've tried the > /proc/sys/vm/(max|min)-readahead setting in Linux without any success. I > need somewhere between 4 and 16 MB readahead per file to be able to read > a large number of large files concurrently, and to reduce seeks to a > minimum. The XFS read path is really the general linux read path, so the readahead logic is the same for both. Take a look in mm/filemap.c looks for MAX_READAHEAD - it is 31 pages, or 124K bytes, you could probably change the algorithm and bump the maximum in here - however, readahead just queues I/O to the block layer, the block layer has its own maximum sizes for individual I/O requests. You may find upping the readahead limit just pushes the problem down the stack so to speak. Steve > > thanks > > roy > > -- > Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA > > Computers are like air conditioners. > They stop working when you open Windows. -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 10:19:48 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5IJmO18886 for linux-xfs-outgoing; Wed, 5 Dec 2001 10:19:48 -0800 Received: from mustard.heime.net (mustard.heime.net [194.234.65.222]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5IJio18864 for ; Wed, 5 Dec 2001 10:19:44 -0800 Received: from localhost (roy@localhost) by mustard.heime.net (8.11.6/8.9.3) with ESMTP id fB5HJNv02720; Wed, 5 Dec 2001 18:19:23 +0100 Date: Wed, 5 Dec 2001 18:19:23 +0100 (CET) From: Roy Sigurd Karlsbakk X-Sender: To: Steve Lord cc: XFS Mailing list Subject: Re: XFS readahead setting??? In-Reply-To: <1007572510.22679.14.camel@jen.americas.sgi.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk > The XFS read path is really the general linux read path, so the > readahead logic is the same for both. Take a look in mm/filemap.c > looks for MAX_READAHEAD - it is 31 pages, or 124K bytes, you could > probably change the algorithm and bump the maximum in here - however, > readahead just queues I/O to the block layer, the block layer has > its own maximum sizes for individual I/O requests. You may find > upping the readahead limit just pushes the problem down the stack > so to speak. ok. the MAX_READAHEAD is 'int vm_max_readahead = 31' in 2.2.16. I guess this is the same as '/proc/sys/vm/max-readahead' (and equally for min-). What do you think I can do to make Linux make LARGE I/O requests, or at least serialize them so I can reduce the seeks? -- Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA Computers are like air conditioners. They stop working when you open Windows. From owner-linux-xfs@oss.sgi.com Wed Dec 5 10:22:18 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5IMIp19045 for linux-xfs-outgoing; Wed, 5 Dec 2001 10:22:18 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5IMBo19023 for ; Wed, 5 Dec 2001 10:22:11 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id SAA1309758 for ; Wed, 5 Dec 2001 18:22:06 +0100 (CET) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3773603; Wed, 5 Dec 2001 11:20:47 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id LAA50555; Wed, 5 Dec 2001 11:20:46 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5HKQf24261; Wed, 5 Dec 2001 11:20:26 -0600 Subject: Re: Files on XFS not safe?! From: Steve Lord To: Xianglong Yuan Cc: linux-xfs@oss.sgi.com In-Reply-To: <20011205120325.C3446@real-uma.mit.edu> References: <20011205100043.A3437@real-uma.mit.edu> <20011205111421.B3446@real-uma.mit.edu> <1007569713.22718.10.camel@jen.americas.sgi.com> <20011205120325.C3446@real-uma.mit.edu> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 11:20:26 -0600 Message-Id: <1007572826.22679.16.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 11:03, Xianglong Yuan wrote: > > I presume there is a log that has the information when the > data-flashing is scheduled to start, whether it has done or not, > and what the original file state (meta-data?) is. Now, if the > system crashed before the data-flashing is done, could it be > possible to trace the log back to the point before the unfinished > data-flashing and give the original file state (meta-data?) back > instead of the meta-data of the new file as it gives nothing? (I > believe the content of the old file is still on the disk and not > zeroed by other program, and could be recovered if we know its > locations (meta-data?).) Update the new meta-data only after a > successful data-flashing. Am I missing something here? Writing file data out to disk in XFS does not go anywhere near the log, only the allocation of extents goes in the log. Also, think about what the editor is probably doing: o Create a new file with a temporary name o write all the file data into it. o remove the original file o rename the temporary file to the orginal name How can the filesystem associate the two files with each other? It is also this rename of the original file which is probably pushing the inode out to disk with the new size. There must also be some operation in there after the rename which is removing a file - and this is a synchronous transaction - which is why the log is on disk at all. Also, once we have removed extents from a file, it is very difficult to work out which ones they were - and the xfs log is of fairly small size, so there is always the possibility that the info is not even in the log anymore. We are working on removing the synchronous transactions, and this would mean that there would be a chance the log write would not have been issued to disk in this case and all we would see would be the original file - as in the ext2 case where all of the pending metadata updates we probably still in memory. Steve > > Xianglong > > -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 10:29:15 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5ITFM19297 for linux-xfs-outgoing; Wed, 5 Dec 2001 10:29:15 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5ITAo19275 for ; Wed, 5 Dec 2001 10:29:10 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id JAA02150 for ; Wed, 5 Dec 2001 09:29:11 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3776024; Wed, 5 Dec 2001 11:27:50 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id LAA40123; Wed, 5 Dec 2001 11:27:50 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5HRTx24275; Wed, 5 Dec 2001 11:27:29 -0600 Subject: Re: XFS readahead setting??? From: Steve Lord To: Roy Sigurd Karlsbakk Cc: XFS Mailing list In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 11:27:29 -0600 Message-Id: <1007573249.22679.20.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 11:19, Roy Sigurd Karlsbakk wrote: > > The XFS read path is really the general linux read path, so the > > readahead logic is the same for both. Take a look in mm/filemap.c > > looks for MAX_READAHEAD - it is 31 pages, or 124K bytes, you could > > probably change the algorithm and bump the maximum in here - however, > > readahead just queues I/O to the block layer, the block layer has > > its own maximum sizes for individual I/O requests. You may find > > upping the readahead limit just pushes the problem down the stack > > so to speak. > > ok. the MAX_READAHEAD is 'int vm_max_readahead = 31' in 2.2.16. I guess > this is the same as '/proc/sys/vm/max-readahead' (and equally for min-). > > What do you think I can do to make Linux make LARGE I/O requests, or at > least serialize them so I can reduce the seeks? And last time this came up, you could not change the application, correct? Otherwise I would say use O_DIRECT in XFS to do the reads, and use threads in the application to do explicit readahead into buffers. Try bumping up the max and see what happens, but I think a lot of the lower layers will automatically split it again, I don't think the IDE code goes bigger than 128K anyway. Steve > > -- > Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA > > Computers are like air conditioners. > They stop working when you open Windows. -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 10:35:49 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5IZnV19530 for linux-xfs-outgoing; Wed, 5 Dec 2001 10:35:49 -0800 Received: from rj.sgi.com (rj.sgi.com [204.94.215.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5IZko19507 for ; Wed, 5 Dec 2001 10:35:46 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by rj.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB5HZdY00554 for ; Wed, 5 Dec 2001 09:35:39 -0800 Received: from thistle-e185.americas.sgi.com (thistle-e185.americas.sgi.com [128.162.185.204]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3612688 for ; Wed, 5 Dec 2001 11:34:23 -0600 (CST) Received: from clink.americas.sgi.com (clink-eth.americas.sgi.com [128.162.2.8]) by thistle-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id LAA39072 for ; Wed, 5 Dec 2001 11:34:22 -0600 (CST) Received: (from roehrich@localhost) by clink.americas.sgi.com (SGI-8.9.3/8.9.3/erikj-IRIX) id LAA31540 for linux-xfs@oss.sgi.com; Wed, 5 Dec 2001 11:34:22 -0600 (CST) Date: Wed, 5 Dec 2001 11:34:22 -0600 (CST) From: Dean Roehrich Message-Id: <200112051734.LAA31540@clink.americas.sgi.com> To: linux-xfs@oss.sgi.com Subject: TAKE - strip nearly all #ifdef __sgi code from dmapi Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Wed Dec 5 09:34:03 PST 2001 Workarea: clink-eth.americas.sgi.com:/data/clink/a67/roehrich/2.4.x-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107818a linux/fs/xfs/xfs_dmapi.c - 1.44 linux/fs/xfs_dmapi/dmapi_session.c - 1.2 linux/fs/xfs_dmapi/dmapi_right.c - 1.2 linux/fs/xfs_dmapi/dmapi_register.c - 1.2 linux/fs/xfs_dmapi/dmapi_private.h - 1.2 linux/fs/xfs_dmapi/dmapi_mountinfo.c - 1.2 linux/fs/xfs_dmapi/dmapi_event.c - 1.2 linux/fs/xfs_dmapi/dmapi_config.c - 1.2 - No Message Supplied From owner-linux-xfs@oss.sgi.com Wed Dec 5 10:48:50 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5ImoK19898 for linux-xfs-outgoing; Wed, 5 Dec 2001 10:48:50 -0800 Received: from mustard.heime.net (mustard.heime.net [194.234.65.222]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Imlo19869 for ; Wed, 5 Dec 2001 10:48:47 -0800 Received: from localhost (roy@localhost) by mustard.heime.net (8.11.6/8.9.3) with ESMTP id fB5HmPP02852; Wed, 5 Dec 2001 18:48:25 +0100 Date: Wed, 5 Dec 2001 18:48:25 +0100 (CET) From: Roy Sigurd Karlsbakk X-Sender: To: Steve Lord cc: XFS Mailing list , Tux mailing list Subject: Re: XFS/Tux readahead setting??? In-Reply-To: <1007573249.22679.20.camel@jen.americas.sgi.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk > And last time this came up, you could not change the application, > correct? Otherwise I would say use O_DIRECT in XFS to do the reads, > and use threads in the application to do explicit readahead into > buffers. > > Try bumping up the max and see what happens, but I think a lot of the > lower layers will automatically split it again, I don't think the IDE > code goes bigger than 128K anyway. I'm currently using the tux web server, and I really don't know how to tell it to read the next - say - 4 megs of the file. I'd be glad if this could be settable in the kernel... -- Roy Sigurd Karlsbakk, MCSE, MCNE, CLS, LCA Computers are like air conditioners. They stop working when you open Windows. From owner-linux-xfs@oss.sgi.com Wed Dec 5 10:49:22 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5InMa20039 for linux-xfs-outgoing; Wed, 5 Dec 2001 10:49:22 -0800 Received: from real-uma.mit.edu (REAL-UMA.MIT.EDU [18.82.0.129]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5InFo20014 for ; Wed, 5 Dec 2001 10:49:15 -0800 Received: (qmail 3476 invoked by uid 1001); 5 Dec 2001 17:49:13 -0000 Date: Wed, 5 Dec 2001 12:49:13 -0500 From: Xianglong Yuan To: linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! Message-ID: <20011205124913.E3446@real-uma.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.3.20i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >-Steve Lord [05 Dec 2001 11:20 -0600] wrote: > > On Wed, 2001-12-05 at 11:03, Xianglong Yuan wrote: > > > > > I presume there is a log that has the information when the > > data-flashing is scheduled to start, whether it has done or not, > > and what the original file state (meta-data?) is. Now, if the > > system crashed before the data-flashing is done, could it be > > possible to trace the log back to the point before the unfinished > > data-flashing and give the original file state (meta-data?) back > > instead of the meta-data of the new file as it gives nothing? (I > > believe the content of the old file is still on the disk and not > > zeroed by other program, and could be recovered if we know its > > locations (meta-data?).) Update the new meta-data only after a > > successful data-flashing. Am I missing something here? > > Writing file data out to disk in XFS does not go anywhere near the log, > only the allocation of extents goes in the log. Also, think about > what the editor is probably doing: > > o Create a new file with a temporary name > o write all the file data into it. > o remove the original file > o rename the temporary file to the orginal name > > How can the filesystem associate the two files with each other? It is > also this rename of the original file which is probably pushing the > inode out to disk with the new size. There must also be some operation > in there after the rename which is removing a file - and this is > a synchronous transaction - which is why the log is on disk at all. > > Also, once we have removed extents from a file, it is very difficult to > work out which ones they were - and the xfs log is of fairly small size, > so there is always the possibility that the info is not even in the > log anymore. > > We are working on removing the synchronous transactions, and this would > mean that there would be a chance the log write would not have been > issued to disk in this case and all we would see would be the original > file - as in the ext2 case where all of the pending metadata updates we > probably still in memory. > > Steve > Thanks, Steve, it becomes much clear to me now as why it is. This problem is critical to us as we run structure simulation for up to several weeks and dump intermediate data every couple hours and append them to few large files. If for some reason the system crash during data output (though rarely), all the results from few weeks running will gone which is unacceptable for us. Probably I should looking for FS journaling both meta-data and file-data, although I don't really need file-data journaling if ever I could get my old content back. Xianglong -- Dept. Materials Science and Engineering MIT, Room 13-4050 Cambridge, MA 02139 From owner-linux-xfs@oss.sgi.com Wed Dec 5 11:04:19 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5J4Jr20503 for linux-xfs-outgoing; Wed, 5 Dec 2001 11:04:19 -0800 Received: from smtpstore.strencom.net ([217.75.0.70]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5J4Go20479 for ; Wed, 5 Dec 2001 11:04:16 -0800 Received: from raptor.raidtec.ie (unknown [217.75.2.18]) by smtpstore.strencom.net (Postfix) with SMTP id 8390863EE2 for ; Mon, 3 Dec 2001 12:00:24 +0000 (GMT) Received: from no.name.available by raptor.raidtec.ie via smtpd (for [217.75.0.68]) with SMTP; 3 Dec 2001 12:06:54 UT content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Subject: Syscall number X-MimeOLE: Produced By Microsoft Exchange V6.0.4712.0 Date: Mon, 3 Dec 2001 12:03:03 -0000 Message-ID: X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: TAKE - xfstests/064 Thread-Index: AcF7wVzTho0pscRjSyaXsUpAzcIc9gAMEoHw From: "Juer Lee" To: Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id fB5J4Go20480 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, All ( XFS team ), We have been testing XFS ACL on Power PC with linux-2.4.5-xfs-1.0.1 for 4 months already, we still haven't found any big problems, so it seems that it is the time to add syscall number for ACL. There are three syscall number which need to be added: __NR__attrctl, __NR__acl_get and __NR__acl_set. Can anybody tell me how to apply them as standard numbers? Thanks. Juer From owner-linux-xfs@oss.sgi.com Wed Dec 5 11:22:18 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5JMID21048 for linux-xfs-outgoing; Wed, 5 Dec 2001 11:22:18 -0800 Received: from zok.sgi.com (zok.SGI.COM [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5JMFo21026 for ; Wed, 5 Dec 2001 11:22:15 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB5IM9A05539 for ; Wed, 5 Dec 2001 10:22:09 -0800 Received: from thistle-e185.americas.sgi.com (thistle-e185.americas.sgi.com [128.162.185.204]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id MAA3626533 for ; Wed, 5 Dec 2001 12:20:53 -0600 (CST) Received: from clink.americas.sgi.com (clink-eth.americas.sgi.com [128.162.2.8]) by thistle-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id MAA85901 for ; Wed, 5 Dec 2001 12:20:52 -0600 (CST) Received: (from roehrich@localhost) by clink.americas.sgi.com (SGI-8.9.3/8.9.3/erikj-IRIX) id MAA08468 for linux-xfs@oss.sgi.com; Wed, 5 Dec 2001 12:20:52 -0600 (CST) Date: Wed, 5 Dec 2001 12:20:52 -0600 (CST) From: Dean Roehrich Message-Id: <200112051820.MAA08468@clink.americas.sgi.com> To: linux-xfs@oss.sgi.com Subject: TAKE - update libdm changelog/version for new dmapi device Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Wed Dec 5 10:19:49 PST 2001 Workarea: clink-eth.americas.sgi.com:/data/clink/a67/roehrich/2.4.x-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107832a cmd/dmapi/libdm/Makefile - 1.6 cmd/dmapi/VERSION - 1.8 cmd/dmapi/debian/changelog - 1.6 cmd/dmapi/doc/CHANGES - 1.8 - No Message Supplied From owner-linux-xfs@oss.sgi.com Wed Dec 5 11:26:48 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5JQmq21214 for linux-xfs-outgoing; Wed, 5 Dec 2001 11:26:48 -0800 Received: from Cantor.suse.de (ns.suse.de [213.95.15.193]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5JQjo21191 for ; Wed, 5 Dec 2001 11:26:45 -0800 Received: from Hermes.suse.de (Hermes.suse.de [213.95.15.136]) by Cantor.suse.de (Postfix) with ESMTP id 9E1291E463; Wed, 5 Dec 2001 19:26:38 +0100 (MET) Date: Wed, 5 Dec 2001 19:26:38 +0100 From: Andi Kleen To: Xianglong Yuan Cc: linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! Message-ID: <20011205192638.A10291@wotan.suse.de> References: <20011205124913.E3446@real-uma.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20011205124913.E3446@real-uma.mit.edu> User-Agent: Mutt/1.3.22.1i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk > Thanks, Steve, it becomes much clear to me now as why it is. This > problem is critical to us as we run structure simulation for up > to several weeks and dump intermediate data every couple hours > and append them to few large files. If for some reason the system > crash during data output (though rarely), all the results from > few weeks running will gone which is unacceptable for us. > Probably I should looking for FS journaling both meta-data and > file-data, although I don't really need file-data journaling if > ever I could get my old content back. The scenario Steve described obviously does not apply to appending to big files, because there is no rename() from a temporary file involved. If it has been flushed they will be on disk. You can force flushing by using a fsync() for example. -Andi From owner-linux-xfs@oss.sgi.com Wed Dec 5 11:35:46 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5JZkJ21497 for linux-xfs-outgoing; Wed, 5 Dec 2001 11:35:46 -0800 Received: from riv-vwsmtpout.echostar.com (wks-253.echostar.com [204.76.128.253]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5JZho21475 for ; Wed, 5 Dec 2001 11:35:44 -0800 Received: from 10.79.98.110 by riv-vwsmtpout.echostar.com (InterScan E-Mail VirusWall NT); Wed, 05 Dec 2001 11:34:50 -0700 Message-ID: <3C0E69EA.A33185CA@echostar.com> Date: Wed, 05 Dec 2001 11:39:38 -0700 From: "Ian S. Nelson" Reply-To: ian.nelson@echostar.com Organization: Echostar X-Mailer: Mozilla 4.73 [en] (X11; U; Linux 2.4.3 i686) X-Accept-Language: en MIME-Version: 1.0 To: "linux-xfs@oss.sgi.com" Subject: Bad write on page x. is this a critical error or a warning? Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk I'm seeing this periodically in my system log. Is this an error or a warning. It looks to me like it needs to flush buffers and free up some pages. I'm streaming a large file at full speed when it happens. thanks, Ian From owner-linux-xfs@oss.sgi.com Wed Dec 5 11:39:35 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5JdZ921674 for linux-xfs-outgoing; Wed, 5 Dec 2001 11:39:35 -0800 Received: from server-3.tower-15.messagelabs.com (mail15.messagelabs.com [63.210.62.243]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5JdRo21650 for ; Wed, 5 Dec 2001 11:39:27 -0800 X-VirusChecked: Checked Received: (qmail 5714 invoked from network); 5 Dec 2001 18:33:39 -0000 Received: from mail.tapeware.com (HELO yt-internet.tapeware.com) (4.21.59.10) by server-3.tower-15.messagelabs.com with SMTP; 5 Dec 2001 18:33:39 -0000 Received: by mail.tapeware.com with Internet Mail Service (5.5.2653.19) id ; Wed, 5 Dec 2001 10:39:18 -0800 Message-ID: From: "Quang Nguyen (Ngo)" To: "'linux-xfs@oss.sgi.com'" Subject: Extended Attributes Date: Wed, 5 Dec 2001 10:39:16 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Disposition: inline Content-Type: text/plain Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk What are the limitations on extended attributes when calling attr_set() and attr_setf()? How much space can each node hold, and what are the side-effects? Does anyone know if other journaling file systems also support extended attributes besides XFS? Thanks, Quang _____________________________________________________________________ This message has been checked for all known viruses by the MessageLabs Virus Scanning Service. For further information visit http://www.messagelabs.com/stats.asp [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Wed Dec 5 11:45:19 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5JjJu21896 for linux-xfs-outgoing; Wed, 5 Dec 2001 11:45:19 -0800 Received: from rj.sgi.com (rj.sgi.com [204.94.215.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5JjFo21874 for ; Wed, 5 Dec 2001 11:45:15 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by rj.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB5Ij9Y06283 for ; Wed, 5 Dec 2001 10:45:09 -0800 Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id MAA3776208; Wed, 5 Dec 2001 12:43:54 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id MAA69478; Wed, 5 Dec 2001 12:43:53 -0600 (CST) Subject: Re: Syscall number From: Eric Sandeen To: Juer Lee Cc: linux-xfs@oss.sgi.com In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 05 Dec 2001 12:37:24 -0600 Message-Id: <1007577444.28030.7.camel@stout.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi Juer - There is an ongoing discussion on LKML about this very subject - some of it has been cross-posted to this list as well. The short answer is that there are not standard numbers available yet, but it's in the works. -Eric On Mon, 2001-12-03 at 06:03, Juer Lee wrote: > Hi, All ( XFS team ), > > We have been testing XFS ACL on Power PC with linux-2.4.5-xfs-1.0.1 for > 4 months already, we still haven't found any big problems, so it seems > that it is the time to add syscall number for ACL. > There are three syscall number which need to be added: __NR__attrctl, > __NR__acl_get and __NR__acl_set. > > Can anybody tell me how to apply them as standard numbers? > > > Thanks. > Juer -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Wed Dec 5 11:46:04 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5Jk4O22039 for linux-xfs-outgoing; Wed, 5 Dec 2001 11:46:04 -0800 Received: from rj.sgi.com (rj.sgi.com [204.94.215.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Jjxo22014 for ; Wed, 5 Dec 2001 11:45:59 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by rj.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB5IjrY06329 for ; Wed, 5 Dec 2001 10:45:53 -0800 Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id MAA3775271; Wed, 5 Dec 2001 12:44:36 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id MAA46161; Wed, 5 Dec 2001 12:44:36 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5IiFH24643; Wed, 5 Dec 2001 12:44:15 -0600 Subject: Re: Extended Attributes From: Steve Lord To: "Quang Nguyen (Ngo)" Cc: "'linux-xfs@oss.sgi.com'" In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 12:44:15 -0600 Message-Id: <1007577855.22718.24.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 12:39, Quang Nguyen (Ngo) wrote: > What are the limitations on extended attributes when calling attr_set() and > attr_setf()? How much space can each node hold, and what are the > side-effects? Each one can hold 64K of data. No side effects really - except it will take a little longer to remove the file. > > Does anyone know if other journaling file systems also support extended > attributes besides XFS? Given there is a patch for ext2, it is not inconceivble that ext3 will support them at some point. Reiserfs will probably end up with them, or something like them, Hans liked multiple streams in a file. Not sure about jfs - it may on os2, but probably does not on Linux. Steve > > Thanks, > Quang > > > _____________________________________________________________________ > This message has been checked for all known viruses by the > MessageLabs Virus Scanning Service. For further information visit > http://www.messagelabs.com/stats.asp > > > [[HTML alternate version deleted]] -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 11:51:10 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5JpAs22296 for linux-xfs-outgoing; Wed, 5 Dec 2001 11:51:10 -0800 Received: from server-3.tower-15.messagelabs.com (mail15.messagelabs.com [63.210.62.243]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Jp5o22243 for ; Wed, 5 Dec 2001 11:51:05 -0800 X-VirusChecked: Checked Received: (qmail 8134 invoked from network); 5 Dec 2001 18:45:16 -0000 Received: from mail.tapeware.com (HELO yt-internet.tapeware.com) (4.21.59.10) by server-3.tower-15.messagelabs.com with SMTP; 5 Dec 2001 18:45:16 -0000 Received: by mail.tapeware.com with Internet Mail Service (5.5.2653.19) id ; Wed, 5 Dec 2001 10:50:56 -0800 Message-ID: From: "Quang Nguyen (Ngo)" To: "'linux-xfs@oss.sgi.com'" Subject: mmap() Limitation Date: Wed, 5 Dec 2001 10:50:54 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Disposition: inline Content-Type: text/plain Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Let's say I want to write files to a tape device using: fd = open(file, O_RDONLY); addr = mmap(0, size, PROT_READ, MAP_PRIVATE, fd, 0); writeTape(addr, size); close(fd); munmap(addr, size); Is there a limit on how big the file be? Should I just use read() and write()? Isn't mmap() faster? There's a paper called "64 Bit File Access" from the XFS site that talks about mmap64(() and stat64(). Should I be using those 64-bit functions instead? Thanks, Quang _____________________________________________________________________ This message has been checked for all known viruses by the MessageLabs Virus Scanning Service. For further information visit http://www.messagelabs.com/stats.asp [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Wed Dec 5 11:51:08 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5Jp8j22288 for linux-xfs-outgoing; Wed, 5 Dec 2001 11:51:08 -0800 Received: from zok.sgi.com (zok.SGI.COM [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Jp5o22242 for ; Wed, 5 Dec 2001 11:51:05 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB5IoxA08287 for ; Wed, 5 Dec 2001 10:50:59 -0800 Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id MAA3773191; Wed, 5 Dec 2001 12:49:42 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id MAA12841; Wed, 5 Dec 2001 12:49:42 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5InK524653; Wed, 5 Dec 2001 12:49:20 -0600 Subject: Re: Bad write on page x. is this a critical error or a warning? From: Steve Lord To: ian.nelson@echostar.com Cc: "linux-xfs@oss.sgi.com" In-Reply-To: <3C0E69EA.A33185CA@echostar.com> References: <3C0E69EA.A33185CA@echostar.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 12:49:20 -0600 Message-Id: <1007578160.22718.26.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 12:39, Ian S. Nelson wrote: > > I'm seeing this periodically in my system log. Is this an error or a > warning. It looks to me like it needs to flush buffers and free up some > pages. I'm streaming a large file at full speed when it happens. The difference between the xfs version of this code and the generic version is the print statement. It has been removed in the latest kernels, so it is not really important. Steve > > thanks, > Ian -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 12:02:09 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5K29s22742 for linux-xfs-outgoing; Wed, 5 Dec 2001 12:02:09 -0800 Received: from dmz.tecosim.de ([194.24.222.241]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5K23o22720 for ; Wed, 5 Dec 2001 12:02:04 -0800 Received: (from uucp@localhost) by dmz.tecosim.de (8.11.0/8.11.0/SuSE Linux 8.11.0-0.4) id fB5J05n08864; Wed, 5 Dec 2001 20:00:05 +0100 Received: from ns.tecosim.de(194.24.222.9) via SMTP by dmz.tecosim.de, id smtpdUFHPK4; Wed Dec 5 20:00:02 2001 Received: from donner.tecosim.de (donner.tecosim.de [194.24.222.109]) by ns.tecosim.de (8.11.3/8.11.3/SuSE Linux 8.11.1-0.5) with ESMTP id fB5J01Z28059; Wed, 5 Dec 2001 20:00:01 +0100 Received: (from leh@localhost) by donner.tecosim.de (8.11.3/8.11.2/SuSE Linux 8.11.1-0.5) id fB5J01622941; Wed, 5 Dec 2001 20:00:01 +0100 Date: Wed, 5 Dec 2001 20:00:01 +0100 From: Utz Lehmann To: "Ian S. Nelson" Cc: "linux-xfs@oss.sgi.com" Subject: Re: Bad write on page x. is this a critical error or a warning? Message-ID: <20011205200001.D1316@de.tecosim.com> References: <3C0E69EA.A33185CA@echostar.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.3.12i In-Reply-To: <3C0E69EA.A33185CA@echostar.com>; from ian.nelson@echostar.com on Wed, Dec 05, 2001 at 11:39:38AM -0700 X-Scanned-By: MIMEDefang 1.3 (www dot roaringpenguin dot com slash mimedefang) Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi I got this error while playing with extremly large files. http://oss.sgi.com/projects/xfs/mail_archive/0110/msg00522.html http://oss.sgi.com/projects/xfs/mail_archive/0110/msg00539.html utz Ian S. Nelson [ian.nelson@echostar.com] wrote: > > I'm seeing this periodically in my system log. Is this an error or a > warning. It looks to me like it needs to flush buffers and free up some > pages. I'm streaming a large file at full speed when it happens. > > thanks, > Ian From owner-linux-xfs@oss.sgi.com Wed Dec 5 12:04:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5K43v22893 for linux-xfs-outgoing; Wed, 5 Dec 2001 12:04:03 -0800 Received: from real-uma.mit.edu (REAL-UMA.MIT.EDU [18.82.0.129]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5K40o22871 for ; Wed, 5 Dec 2001 12:04:00 -0800 Received: (qmail 3486 invoked by uid 1001); 5 Dec 2001 19:03:58 -0000 Date: Wed, 5 Dec 2001 14:03:58 -0500 From: Xianglong Yuan To: linux-xfs@oss.sgi.com Subject: Follow up -- Re: Files on XFS not safe?! Message-ID: <20011205140358.F3446@real-uma.mit.edu> References: <20011205124913.E3446@real-uma.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20011205124913.E3446@real-uma.mit.edu> User-Agent: Mutt/1.3.20i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk I just cross-by an excellent article on various journaling techniques used in ext3. http://www-106.ibm.com/developerworks/linux/library/l-fs8.html There are three modes in ext3: writeback mode, ordered mode, and journal mode. Most journaling FSs use writeback mode which allows possible file corruption as I mentioned in this thread. While journal mode, which provides full data and metadata journaling, is the safest but too expensive and unnecessary for me, ordered mode is exactly what I want: possible loss of updates only and no file corruption. It surprises me that ext3, now in the official 2.4 kernel, provides all these journaling techniques. Folks, any reason against my likely defection? :-)) I recall someone said ext3 is not a true journaling FS. Is it? Xianglong From owner-linux-xfs@oss.sgi.com Wed Dec 5 12:19:34 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5KJYq23359 for linux-xfs-outgoing; Wed, 5 Dec 2001 12:19:34 -0800 Received: from real-uma.mit.edu (REAL-UMA.MIT.EDU [18.82.0.129]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5KJTo23335 for ; Wed, 5 Dec 2001 12:19:29 -0800 Received: (qmail 3491 invoked by uid 1001); 5 Dec 2001 19:19:27 -0000 Date: Wed, 5 Dec 2001 14:19:27 -0500 From: Xianglong Yuan To: Andi Kleen Cc: linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! Message-ID: <20011205141927.G3446@real-uma.mit.edu> References: <20011205124913.E3446@real-uma.mit.edu> <20011205192638.A10291@wotan.suse.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20011205192638.A10291@wotan.suse.de> User-Agent: Mutt/1.3.20i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >-Andi Kleen [05 Dec 2001 19:26 +0100] wrote: > > > Thanks, Steve, it becomes much clear to me now as why it is. This > > problem is critical to us as we run structure simulation for up > > to several weeks and dump intermediate data every couple hours > > and append them to few large files. If for some reason the system > > crash during data output (though rarely), all the results from > > few weeks running will gone which is unacceptable for us. > > Probably I should looking for FS journaling both meta-data and > > file-data, although I don't really need file-data journaling if > > ever I could get my old content back. > > The scenario Steve described obviously does not apply to appending to big > files, because there is no rename() from a temporary file involved. > If it has been flushed they will be on disk. You can force flushing > by using a fsync() for example. > > -Andi But only be on disk after flushing, correct? What if crashing during data dumping before it can issue a fsync(), or even during fsync()? I think the best bet is to remove the old file only after ensuring the new one is there, or fallback to old one, an asynchronous trasaction which is I believe what Steve said he is working on. Same technique as ordered mode in ext3? Xianglong From owner-linux-xfs@oss.sgi.com Wed Dec 5 12:21:20 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5KLK323507 for linux-xfs-outgoing; Wed, 5 Dec 2001 12:21:20 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5KLGo23484 for ; Wed, 5 Dec 2001 12:21:16 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id LAA09429 for ; Wed, 5 Dec 2001 11:21:15 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id NAA3774220; Wed, 5 Dec 2001 13:19:59 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id NAA83488; Wed, 5 Dec 2001 13:19:58 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5JJav29161; Wed, 5 Dec 2001 13:19:36 -0600 Subject: Re: mmap() Limitation From: Steve Lord To: "Quang Nguyen (Ngo)" Cc: "'linux-xfs@oss.sgi.com'" In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 13:19:35 -0600 Message-Id: <1007579975.22679.28.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 12:50, Quang Nguyen (Ngo) wrote: > Let's say I want to write files to a tape device using: > > fd = open(file, O_RDONLY); > addr = mmap(0, size, PROT_READ, MAP_PRIVATE, fd, 0); > writeTape(addr, size); > close(fd); > munmap(addr, size); > Is there a limit on how big the file be? > > Should I just use read() and write()? Isn't mmap() faster? > > There's a paper called "64 Bit File Access" from the XFS site that talks > about mmap64(() and stat64(). Should I be using those 64-bit functions > instead? There does not appear to be a user space interface to mmap out beyond a 4G boundary on ia32 linux. The mmap64 call is from irix. This looks like a deficiency in the linux large file support to me. Steve -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 12:24:45 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5KOje23701 for linux-xfs-outgoing; Wed, 5 Dec 2001 12:24:45 -0800 Received: from Cantor.suse.de (ns.suse.de [213.95.15.193]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5KOdo23679 for ; Wed, 5 Dec 2001 12:24:39 -0800 Received: from Hermes.suse.de (Hermes.suse.de [213.95.15.136]) by Cantor.suse.de (Postfix) with ESMTP id 3B98F1E190; Wed, 5 Dec 2001 20:24:33 +0100 (MET) Date: Wed, 5 Dec 2001 20:24:33 +0100 From: Andi Kleen To: Xianglong Yuan Cc: Andi Kleen , linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! Message-ID: <20011205202433.A29508@wotan.suse.de> References: <20011205124913.E3446@real-uma.mit.edu> <20011205192638.A10291@wotan.suse.de> <20011205141927.G3446@real-uma.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20011205141927.G3446@real-uma.mit.edu> User-Agent: Mutt/1.3.22.1i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 02:19:27PM -0500, Xianglong Yuan wrote: > >-Andi Kleen [05 Dec 2001 19:26 +0100] wrote: > > > > > Thanks, Steve, it becomes much clear to me now as why it is. This > > > problem is critical to us as we run structure simulation for up > > > to several weeks and dump intermediate data every couple hours > > > and append them to few large files. If for some reason the system > > > crash during data output (though rarely), all the results from > > > few weeks running will gone which is unacceptable for us. > > > Probably I should looking for FS journaling both meta-data and > > > file-data, although I don't really need file-data journaling if > > > ever I could get my old content back. > > > > The scenario Steve described obviously does not apply to appending to big > > files, because there is no rename() from a temporary file involved. > > If it has been flushed they will be on disk. You can force flushing > > by using a fsync() for example. > > > > -Andi > > But only be on disk after flushing, correct? What if crashing Yes, but until that the old data is not gone because the file is not truncated . The problem with the editors is that the old data is usually lost in some temporary file; that won't be the problem here. > during data dumping before it can issue a fsync(), or even during > fsync()? I think the best bet is to remove the old file only Then you lose a few seconds at worst. > after ensuring the new one is there, or fallback to old one, an > asynchronous trasaction which is I believe what Steve said he is > working on. Same technique as ordered mode in ext3? The problem is really that the delete part of rename is synchronous, which causes a log flush at an very inconvenient moment if you're using an editor and crashed shortly afterwards. For appending to existing files it is no problem though again (first no delete and even if there was a flush it wouldn't truncate your file) -Andi From owner-linux-xfs@oss.sgi.com Wed Dec 5 12:25:43 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5KPhv23850 for linux-xfs-outgoing; Wed, 5 Dec 2001 12:25:43 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5KPYo23826 for ; Wed, 5 Dec 2001 12:25:34 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id LAA04412 for ; Wed, 5 Dec 2001 11:25:26 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id NAA3774139; Wed, 5 Dec 2001 13:24:10 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id NAA88579; Wed, 5 Dec 2001 13:24:10 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5JNm429169; Wed, 5 Dec 2001 13:23:48 -0600 Subject: Re: Files on XFS not safe?! From: Steve Lord To: Xianglong Yuan Cc: Andi Kleen , linux-xfs@oss.sgi.com In-Reply-To: <20011205141927.G3446@real-uma.mit.edu> References: <20011205124913.E3446@real-uma.mit.edu> <20011205192638.A10291@wotan.suse.de> <20011205141927.G3446@real-uma.mit.edu> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 13:23:48 -0600 Message-Id: <1007580228.22679.30.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 13:19, Xianglong Yuan wrote: > >-Andi Kleen [05 Dec 2001 19:26 +0100] wrote: > > > > > Thanks, Steve, it becomes much clear to me now as why it is. This > > > problem is critical to us as we run structure simulation for up > > > to several weeks and dump intermediate data every couple hours > > > and append them to few large files. If for some reason the system > > > crash during data output (though rarely), all the results from > > > few weeks running will gone which is unacceptable for us. > > > Probably I should looking for FS journaling both meta-data and > > > file-data, although I don't really need file-data journaling if > > > ever I could get my old content back. > > > > The scenario Steve described obviously does not apply to appending to big > > files, because there is no rename() from a temporary file involved. > > If it has been flushed they will be on disk. You can force flushing > > by using a fsync() for example. > > > > -Andi > > But only be on disk after flushing, correct? What if crashing > during data dumping before it can issue a fsync(), or even during > fsync()? I think the best bet is to remove the old file only > after ensuring the new one is there, or fallback to old one, an > asynchronous trasaction which is I believe what Steve said he is > working on. Same technique as ordered mode in ext3? You can also use O_SYNC when you open the file, and with xfs the file data and associated metadata will be on disk before the write call returns. There is a mount option osyncisdsync which makes go faster by not transactionally updating file modification times. Performance will be disk bound, but if you are moving large amounts of data you may well be disk bound anyway. The ordered mode in ext3 is almost certainly a better performing solution than this, it would be nice to get this into xfs, but probably quite a while before it could be worked on. Steve > > Xianglong -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 12:49:29 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5KnTS24314 for linux-xfs-outgoing; Wed, 5 Dec 2001 12:49:29 -0800 Received: from Cantor.suse.de (ns.suse.de [213.95.15.193]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5KnPo24292 for ; Wed, 5 Dec 2001 12:49:26 -0800 Received: from Hermes.suse.de (Hermes.suse.de [213.95.15.136]) by Cantor.suse.de (Postfix) with ESMTP id 745841E26C; Wed, 5 Dec 2001 20:49:19 +0100 (MET) Date: Wed, 5 Dec 2001 20:49:17 +0100 From: Andi Kleen To: Steve Lord Cc: "Quang Nguyen (Ngo)" , "'linux-xfs@oss.sgi.com'" Subject: Re: mmap() Limitation Message-ID: <20011205204917.A5129@wotan.suse.de> References: <1007579975.22679.28.camel@jen.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1007579975.22679.28.camel@jen.americas.sgi.com> User-Agent: Mutt/1.3.22.1i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk > There does not appear to be a user space interface to mmap out beyond > a 4G boundary on ia32 linux. The mmap64 call is from irix. This looks > like a deficiency in the linux large file support to me. There is actually. In kernel it is called mmap2(), in glibc it should be mmap64() or even default when compiled for largefiles. mmap2 does not pass in a true 64bit offset, but an 32bit scaled by PAGE_SIZE, but that works well enough as the page cache cannot represent bigger files anyways (that is the internal interface between glibc and kernel; the user passes an byte offset to mmap64 of course) -andi From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:25:51 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5LPpo25070 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:25:51 -0800 Received: from UberGeek ([209.184.141.163]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5LPgo25045 for ; Wed, 5 Dec 2001 13:25:42 -0800 Received: (qmail 3762 invoked by uid 500); 5 Dec 2001 20:24:09 -0000 Subject: Re: Follow up -- Re: Files on XFS not safe?! From: Austin Gonyou To: Xianglong Yuan Cc: linux-xfs@oss.sgi.com In-Reply-To: <20011205140358.F3446@real-uma.mit.edu> References: <20011205140358.F3446@real-uma.mit.edu> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-TImXZbG/B7ucIH41Sk2P" X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 14:24:09 -0600 Message-Id: <1007583849.3676.2.camel@UberGeek> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk --=-TImXZbG/B7ucIH41Sk2P Content-Type: text/plain Content-Transfer-Encoding: quoted-printable EXT3 seems to be a "true" journalling FS now. At one point in time though it was kludgy. Now EXT3 is definitely a good journaling filesystem, but it's speed, if you need it, does leave something to be desired. It's best to use IOzone or the mongo.pl benchmarks though to see for yourself in the environment you're going to run. YMMV per FS/data type. On Wed, 2001-12-05 at 13:03, Xianglong Yuan wrote: > I just cross-by an excellent article on various journaling > techniques used in ext3. >=20 > http://www-106.ibm.com/developerworks/linux/library/l-fs8.html >=20 > There are three modes in ext3: writeback mode, ordered mode, and > journal mode. Most journaling FSs use writeback mode which > allows possible file corruption as I mentioned in this thread. > While journal mode, which provides full data and metadata > journaling, is the safest but too expensive and unnecessary for > me, ordered mode is exactly what I want: possible loss of updates > only and no file corruption. It surprises me that ext3, now in > the official 2.4 kernel, provides all these journaling > techniques. Folks, any reason against my likely defection? :-)) > I recall someone said ext3 is not a true journaling FS. Is it? >=20 > Xianglong --=20 Austin Gonyou Systems Architect, CCNA=20 Coremetrics, Inc. Phone: 512-698-7250 email: austin@coremetrics.com =20 "Be ashamed to die until you have won some victory for humanity." Horace Mann, address at Antioch College, 1859 --=-TImXZbG/B7ucIH41Sk2P Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQA8DoJp94g6ZVmFMoIRAlLPAJ9F9LO9UePbGWNBDeXZ4+kZv65nUACfQHIO OVqvlzQb7d0vrfu2Lr5iQPc= =kf8f -----END PGP SIGNATURE----- --=-TImXZbG/B7ucIH41Sk2P-- From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:27:50 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5LRop26066 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:27:50 -0800 Received: from UberGeek ([209.184.141.163]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5LRio26042 for ; Wed, 5 Dec 2001 13:27:45 -0800 Received: (qmail 3775 invoked by uid 500); 5 Dec 2001 20:26:03 -0000 Subject: XFS ACL implementation. From: Austin Gonyou To: "Quang Nguyen \(Ngo\)" Cc: "'linux-xfs@oss.sgi.com'" In-Reply-To: References: Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-gmHdb8Q31cXOGusHQFWP" X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 14:26:03 -0600 Message-Id: <1007583963.3687.4.camel@UberGeek> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk --=-gmHdb8Q31cXOGusHQFWP Content-Type: text/plain Content-Transfer-Encoding: quoted-printable I have a question on the ACL implementation in XFS, is it possible to have "roles" setup so certain people can access certain filesystems, etc, regardless of the higher group perms?=20 --=20 Austin Gonyou Systems Architect, CCNA=20 Coremetrics, Inc. Phone: 512-698-7250 email: austin@coremetrics.com =20 "Be ashamed to die until you have won some victory for humanity." Horace Mann, address at Antioch College, 1859 --=-gmHdb8Q31cXOGusHQFWP Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQA8DoLa94g6ZVmFMoIRAm3dAKCAmdrabOYuWkrLKFW1/rH3iZj4/ACg7NX2 gvCUlVqRkp09IknQT0dy0M4= =H17B -----END PGP SIGNATURE----- --=-gmHdb8Q31cXOGusHQFWP-- From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:37:40 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5Lben27956 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:37:40 -0800 Received: from atlrel7.hp.com (atlrel7.hp.com [156.153.255.213]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Lbbo27933 for ; Wed, 5 Dec 2001 13:37:37 -0800 Received: from xatlrelay1.atl.hp.com (unknown [15.45.89.190]) by atlrel7.hp.com (Postfix) with ESMTP id 048F91F54C for ; Wed, 5 Dec 2001 15:34:15 -0500 (EST) Received: from sa-bwmail1.esr.hp.com (sa-bwmail1.esr.hp.com [15.1.192.32]) by xatlrelay1.atl.hp.com (Postfix) with ESMTP id 992E61F507 for ; Wed, 5 Dec 2001 15:37:25 -0500 (EST) Received: by sa-bwmail1.esr.hp.com with Internet Mail Service (5.5.2653.19) id ; Wed, 5 Dec 2001 15:36:57 -0500 Message-ID: <23D04BDBA646D411BDDD00D0B774B539046029AE@sa-bwmail1.esr.hp.com> From: "Christian, Chip" To: "Linux XFS (E-mail)" Subject: compilers Date: Wed, 5 Dec 2001 15:36:57 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Just to add another data point... I recently started using the 2.4.9-12 kernel on a Dell 2450 with an Intel e1000 gig-ethernet card running RedHat Linux 7.1. The SGI-built kernels work. The RedHat-built kernels work. When I built my own kernels, the module would load, the card would get ifconfig'd up, etc. Couldn't pass traffic. Verified by connecting two identical machines with identical kernels via fibre, which is to say I don't know if traffic went in but not out, out but not in, or neither. Same kernel works just fine on a Dell 6450 with the same e1000 card. I switched CC from kgcc to gcc-2.96-85, made no other changes, and it now works great. From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:40:12 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5LeCC28907 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:40:12 -0800 Received: from rcpt-expgw.biglobe.ne.jp (rcpt-expgw.biglobe.ne.jp [210.147.6.212]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Le6o28884 for ; Wed, 5 Dec 2001 13:40:06 -0800 Received: from smtp-gw.biglobe.ne.jp by rcpt-expgw.biglobe.ne.jp (8.9.3+3.2W/3.7W-01111312) with ESMTP id FAA20586 for ; Thu, 6 Dec 2001 05:40:09 +0900 (JST) X-Biglobe-Sender: Received: from feldmark01 (211.99.247.66 [211.99.247.66]) by smtp-gw.biglobe.ne.jp id FCJIC0A82618; Thu, 06 Dec 2001 05:40:03 +0900 (JST) Message-ID: <02a301c17dcc$71d79a00$3a0a010a@feldmark01> From: "Mark Feldman" To: References: <1007577855.22718.24.camel@jen.americas.sgi.com> Subject: Re: Extended Attributes Date: Thu, 6 Dec 2001 05:35:55 +0900 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-2022-jp" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2919.6700 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk How about limitations on the total number of EA's per file or the total space they can take up per file. Could I have 200, 64K EA's per file? 500? 1000? 10000? Mark ----- Original Message ----- From: "Steve Lord" To: "Quang Nguyen (Ngo)" Cc: Sent: Thursday, December 06, 2001 3:44 AM Subject: Re: Extended Attributes > On Wed, 2001-12-05 at 12:39, Quang Nguyen (Ngo) wrote: > > What are the limitations on extended attributes when calling attr_set() and > > attr_setf()? How much space can each node hold, and what are the > > side-effects? > > Each one can hold 64K of data. No side effects really - except it will > take a little longer to remove the file. > > > > > Does anyone know if other journaling file systems also support extended > > attributes besides XFS? > > Given there is a patch for ext2, it is not inconceivble that ext3 > will support them at some point. Reiserfs will probably end up with > them, or something like them, Hans liked multiple streams in a file. > Not sure about jfs - it may on os2, but probably does not on Linux. > > Steve > > > > > Thanks, > > Quang > > > > > > _____________________________________________________________________ > > This message has been checked for all known viruses by the > > MessageLabs Virus Scanning Service. For further information visit > > http://www.messagelabs.com/stats.asp > > > > > > [[HTML alternate version deleted]] > -- > > Steve Lord voice: +1-651-683-3511 > Principal Engineer, Filesystem Software email: lord@sgi.com > From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:41:31 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5LfVK29064 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:41:31 -0800 Received: from guardian.hermes.si (guardian.hermes.si [193.77.5.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5LfOo29039 for ; Wed, 5 Dec 2001 13:41:25 -0800 Received: from hermes.si (primus.hermes.si [193.77.5.98]) by guardian.hermes.si (8.9.3/8.9.3) with ESMTP id VAA16259; Wed, 5 Dec 2001 21:40:25 +0100 (MET) Received: (from root@localhost) by hermes.si (8.9.3/8.9.3) id VAA07330; Wed, 5 Dec 2001 21:40:23 +0100 Received: from hal9000.hermes.si (hal9000.hermes.si [10.17.5.136]) by hermes.si (8.9.3/8.9.3) with ESMTP id VAA07240; Wed, 5 Dec 2001 21:40:22 +0100 Received: by hal9000.hermes.si with Internet Mail Service (5.5.2650.21) id ; Wed, 5 Dec 2001 21:40:23 +0100 Message-ID: From: Igor Lautar To: "'Christian, Chip'" , "Linux XFS (E-mail)" Subject: RE: compilers Date: Wed, 5 Dec 2001 21:40:21 +0100 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain X-AntiVirus: scanned for viruses by AMaViS 0.2.1 (http://amavis.org/) Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk The default compiler in RH7.1 was broken... I've also had a few problems with it myself... Igor > -----Original Message----- > From: Christian, Chip [mailto:chip_christian@hp.com] > Sent: 5. december 2001 21:37 > To: Linux XFS (E-mail) > Subject: compilers > > > Just to add another data point... > > I recently started using the 2.4.9-12 kernel on a Dell 2450 with an > Intel e1000 gig-ethernet card running RedHat Linux 7.1. > > The SGI-built kernels work. > The RedHat-built kernels work. > When I built my own kernels, the module would load, > the card would get ifconfig'd up, etc. Couldn't pass traffic. > > Verified by connecting two identical machines with identical > kernels via fibre, which is to say I don't know if traffic > went in but not out, out but not in, or neither. > > Same kernel works just fine on a Dell 6450 with the same > e1000 card. > > I switched CC from kgcc to gcc-2.96-85, made no other > changes, and it now works great. > > From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:45:42 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5Ljg530352 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:45:42 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Ljbo30322 for ; Wed, 5 Dec 2001 13:45:37 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id MAA03081 for ; Wed, 5 Dec 2001 12:45:38 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id OAA3546486; Wed, 5 Dec 2001 14:44:18 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id OAA87325; Wed, 5 Dec 2001 14:44:18 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5Khub32069; Wed, 5 Dec 2001 14:43:56 -0600 Subject: Re: compilers From: Steve Lord To: "Christian, Chip" Cc: "Linux XFS (E-mail)" In-Reply-To: <23D04BDBA646D411BDDD00D0B774B539046029AE@sa-bwmail1.esr.hp.com> References: <23D04BDBA646D411BDDD00D0B774B539046029AE@sa-bwmail1.esr.hp.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 14:43:56 -0600 Message-Id: <1007585036.30179.2.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 14:36, Christian, Chip wrote: > Just to add another data point... > > I recently started using the 2.4.9-12 kernel on a Dell 2450 with an > Intel e1000 gig-ethernet card running RedHat Linux 7.1. > > The SGI-built kernels work. > The RedHat-built kernels work. > When I built my own kernels, the module would load, > the card would get ifconfig'd up, etc. Couldn't pass traffic. > > Verified by connecting two identical machines with identical > kernels via fibre, which is to say I don't know if traffic went in > but not out, out but not in, or neither. > > Same kernel works just fine on a Dell 6450 with the same > e1000 card. > > I switched CC from kgcc to gcc-2.96-85, made no other > changes, and it now works great. I am using gcc-2.96-101 which is from the latest rawhide, this is supposed to contain the fix for the problem which has bitten xfs from time to time. I cannot really speak for it too much yet since I am mostly working on 2.5 right now and it does not exactly boot all the way up reliably yet. Steve > -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:46:12 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5LkCE30480 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:46:12 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Lk8o30458 for ; Wed, 5 Dec 2001 13:46:08 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id MAA04811 for ; Wed, 5 Dec 2001 12:46:09 -0800 (PST) mail_from (sandeen@sgi.com) Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id OAA3776037; Wed, 5 Dec 2001 14:44:51 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id OAA77333; Wed, 5 Dec 2001 14:44:50 -0600 (CST) Subject: Re: compilers From: Eric Sandeen To: "Christian, Chip" Cc: "Linux XFS (E-mail)" In-Reply-To: <23D04BDBA646D411BDDD00D0B774B539046029AE@sa-bwmail1.esr.hp.com> References: <23D04BDBA646D411BDDD00D0B774B539046029AE@sa-bwmail1.esr.hp.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 05 Dec 2001 14:38:20 -0600 Message-Id: <1007584700.27980.24.camel@stout.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 14:36, Christian, Chip wrote: > The SGI-built kernels work. > The RedHat-built kernels work. > When I built my own kernels, the module would load, > the card would get ifconfig'd up, etc. Couldn't pass traffic. ... > I switched CC from kgcc to gcc-2.96-85, made no other > changes, and it now works great. Which is even more interesting since: The SGI-built kernels used kgcc The RedHat-built kernels used gcc-2.96-85 AFAIK :) Although something about your config could have perturbed the code to expose a compiler bug... -Eric -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:46:47 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5LklZ30611 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:46:47 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Lkeo30589 for ; Wed, 5 Dec 2001 13:46:40 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id MAA00935 for ; Wed, 5 Dec 2001 12:46:41 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id OAA3770753; Wed, 5 Dec 2001 14:45:22 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id OAA09605; Wed, 5 Dec 2001 14:45:22 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5KixG32073; Wed, 5 Dec 2001 14:44:59 -0600 Subject: Re: Extended Attributes From: Steve Lord To: Mark Feldman Cc: linux-xfs@oss.sgi.com In-Reply-To: <02a301c17dcc$71d79a00$3a0a010a@feldmark01> References: <1007577855.22718.24.camel@jen.americas.sgi.com> <02a301c17dcc$71d79a00$3a0a010a@feldmark01> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 14:44:59 -0600 Message-Id: <1007585099.30459.4.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 14:35, Mark Feldman wrote: > How about limitations on the total number of EA's per file or the total > space they can take up per file. Could I have 200, 64K EA's per file? 500? > 1000? 10000? No practical limit - they all need different names of course, the organization of the entries is almost identical to directories. I cannot say we do extensive testing of vast attribute spaces, there tend never to be more than a handful at the most. Steve > > Mark > > ----- Original Message ----- > From: "Steve Lord" > To: "Quang Nguyen (Ngo)" > Cc: > Sent: Thursday, December 06, 2001 3:44 AM > Subject: Re: Extended Attributes > > > > On Wed, 2001-12-05 at 12:39, Quang Nguyen (Ngo) wrote: > > > What are the limitations on extended attributes when calling attr_set() > and > > > attr_setf()? How much space can each node hold, and what are the > > > side-effects? > > > > Each one can hold 64K of data. No side effects really - except it will > > take a little longer to remove the file. > > > > > > > > Does anyone know if other journaling file systems also support extended > > > attributes besides XFS? > > > > Given there is a patch for ext2, it is not inconceivble that ext3 > > will support them at some point. Reiserfs will probably end up with > > them, or something like them, Hans liked multiple streams in a file. > > Not sure about jfs - it may on os2, but probably does not on Linux. > > > > Steve > > > > > > > > Thanks, > > > Quang > > > > > > > > > _____________________________________________________________________ > > > This message has been checked for all known viruses by the > > > MessageLabs Virus Scanning Service. For further information visit > > > http://www.messagelabs.com/stats.asp > > > > > > > > > [[HTML alternate version deleted]] > > -- > > > > Steve Lord voice: +1-651-683-3511 > > Principal Engineer, Filesystem Software email: lord@sgi.com > > -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:47:27 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5LlR530740 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:47:27 -0800 Received: from real-uma.mit.edu (REAL-UMA.MIT.EDU [18.82.0.129]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5LlLo30718 for ; Wed, 5 Dec 2001 13:47:21 -0800 Received: (qmail 3510 invoked by uid 1001); 5 Dec 2001 20:47:18 -0000 Date: Wed, 5 Dec 2001 15:47:18 -0500 From: Xianglong Yuan To: linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! Message-ID: <20011205154718.A3506@real-uma.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.3.20i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >-Andi Kleen [05 Dec 2001 20:24 +0100] wrote: > > On Wed, Dec 05, 2001 at 02:19:27PM -0500, Xianglong Yuan wrote: > > >-Andi Kleen [05 Dec 2001 19:26 +0100] wrote: > > > > > > > Thanks, Steve, it becomes much clear to me now as why it is. This > > > > problem is critical to us as we run structure simulation for up > > > > to several weeks and dump intermediate data every couple hours > > > > and append them to few large files. If for some reason the system > > > > crash during data output (though rarely), all the results from > > > > few weeks running will gone which is unacceptable for us. > > > > Probably I should looking for FS journaling both meta-data and > > > > file-data, although I don't really need file-data journaling if > > > > ever I could get my old content back. > > > > > > The scenario Steve described obviously does not apply to appending to big > > > files, because there is no rename() from a temporary file involved. > > > If it has been flushed they will be on disk. You can force flushing > > > by using a fsync() for example. > > > > > > -Andi > > > > But only be on disk after flushing, correct? What if crashing > > Yes, but until that the old data is not gone because the file is not truncated . > The problem with the editors is that the old data is usually lost in some > temporary file; that won't be the problem here. > > > during data dumping before it can issue a fsync(), or even during > > fsync()? I think the best bet is to remove the old file only > > Then you lose a few seconds at worst. > > > after ensuring the new one is there, or fallback to old one, an > > asynchronous trasaction which is I believe what Steve said he is > > working on. Same technique as ordered mode in ext3? > > The problem is really that the delete part of rename is synchronous, > which causes a log flush at an very inconvenient moment if you're using an editor > and crashed shortly afterwards. For appending to existing files it is no problem > though again (first no delete and even if there was a flush it wouldn't truncate > your file) > > > -Andi > I see what I am testing using vim is different from what my simulation actually recording data appendingly :-<< But, anyhow, ordered mode in ext3 seems to be my best bet. I'll do some reading to find out the cons in ext3 :-)) Any comments on ext3 will be highly welcomed. Please send me privately if off-topic. Thanks. Xianglong From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:52:02 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5Lq2K32393 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:52:02 -0800 Received: from UberGeek ([209.184.141.163]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Lpro32369 for ; Wed, 5 Dec 2001 13:51:53 -0800 Received: (qmail 2087 invoked by uid 500); 5 Dec 2001 20:50:21 -0000 Subject: RE: compilers From: Austin Gonyou To: Igor Lautar Cc: "'Christian, Chip'" , "Linux XFS \(E-mail\)" In-Reply-To: References: Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-f5HJlfElfSSUrBz7A6eZ" X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 14:50:20 -0600 Message-Id: <1007585421.2026.0.camel@UberGeek> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk --=-f5HJlfElfSSUrBz7A6eZ Content-Type: text/plain Content-Transfer-Encoding: quoted-printable This is not true. RH 7.1 gcc is not broke(as in can't compile code that runs), but there was an update to it which fixed certain issues. You definitely CAN compile kernels with that default gcc though that will boot. Will you have problems? ymmv. The best course of action is to either get gcc3, or update your current gcc from updates.redhat.com. On Wed, 2001-12-05 at 14:40, Igor Lautar wrote: > The default compiler in RH7.1 was broken... > I've also had a few problems with it myself... >=20 > Igor >=20 > > -----Original Message----- > > From: Christian, Chip [mailto:chip_christian@hp.com]=20 > > Sent: 5. december 2001 21:37 > > To: Linux XFS (E-mail) > > Subject: compilers > >=20 > >=20 > > Just to add another data point... > >=20 > > I recently started using the 2.4.9-12 kernel on a Dell 2450 with an=20 > > Intel e1000 gig-ethernet card running RedHat Linux 7.1. > >=20 > > The SGI-built kernels work.=20=20 > > The RedHat-built kernels work.=20=20 > > When I built my own kernels, the module would load,=20 > > the card would get ifconfig'd up, etc. Couldn't pass traffic.=20=20 > >=20 > > Verified by connecting two identical machines with identical=20 > > kernels via fibre, which is to say I don't know if traffic=20 > > went in but not out, out but not in, or neither.=20=20 > >=20 > > Same kernel works just fine on a Dell 6450 with the same=20 > > e1000 card.=20=20 > >=20 > > I switched CC from kgcc to gcc-2.96-85, made no other=20 > > changes, and it now works great. > >=20 > >=20 --=20 Austin Gonyou Systems Architect, CCNA=20 Coremetrics, Inc. Phone: 512-698-7250 email: austin@coremetrics.com =20 "Be ashamed to die until you have won some victory for humanity." Horace Mann, address at Antioch College, 1859 --=-f5HJlfElfSSUrBz7A6eZ Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQA8DoiM94g6ZVmFMoIRAqCyAJ9qg+nePgMSYpMxi9HGr4zAhj+aLQCcCMWS ArrRrSlzxWHMYnm/ZEa93js= =W8Sn -----END PGP SIGNATURE----- --=-f5HJlfElfSSUrBz7A6eZ-- From owner-linux-xfs@oss.sgi.com Wed Dec 5 13:56:15 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5LuFo32554 for linux-xfs-outgoing; Wed, 5 Dec 2001 13:56:15 -0800 Received: from guardian.hermes.si (guardian.hermes.si [193.77.5.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5LuAo32529 for ; Wed, 5 Dec 2001 13:56:11 -0800 Received: from hermes.si (primus.hermes.si [193.77.5.98]) by guardian.hermes.si (8.9.3/8.9.3) with ESMTP id VAA16843 for ; Wed, 5 Dec 2001 21:56:02 +0100 (MET) Received: (from root@localhost) by hermes.si (8.9.3/8.9.3) id VAA25989 for linux-xfs@oss.sgi.com; Wed, 5 Dec 2001 21:56:00 +0100 Received: from hal9000.hermes.si (hal9000.hermes.si [10.17.5.136]) by hermes.si (8.9.3/8.9.3) with ESMTP id VAA25832 for ; Wed, 5 Dec 2001 21:55:58 +0100 Received: by hal9000.hermes.si with Internet Mail Service (5.5.2650.21) id ; Wed, 5 Dec 2001 21:55:58 +0100 Message-ID: From: Igor Lautar To: "Linux XFS (E-mail)" Subject: RE: compilers Date: Wed, 5 Dec 2001 21:55:55 +0100 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain X-AntiVirus: scanned for viruses by AMaViS 0.2.1 (http://amavis.org/) Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk > -----Original Message----- > From: Austin Gonyou [mailto:austin@coremetrics.com] > Sent: 5. december 2001 21:50 > To: Igor Lautar > Cc: 'Christian, Chip'; Linux XFS (E-mail) > Subject: RE: compilers > > > This is not true. RH 7.1 gcc is not broke(as in can't compile > code that runs), but there was an update to it which fixed > certain issues. You definitely CAN compile kernels with that > default gcc though that will boot. Will you have problems? I was refering to that, just haven't been clear enough... I had software that couldn't compile on gcc that comes with rh7.1 ( that's default gcc, not updated ). I think one of them was mplayer ( but not 100% sure ). > ymmv. The best course of action is to either get gcc3, or > update your current gcc from updates.redhat.com. Now I use slackware 8 :) just upgraded to 2.4.16 with xfs support and it works great... Igor From owner-linux-xfs@oss.sgi.com Wed Dec 5 14:01:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5M13100366 for linux-xfs-outgoing; Wed, 5 Dec 2001 14:01:03 -0800 Received: from mail.dkp.com ([204.191.16.3]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5M0xo00338 for ; Wed, 5 Dec 2001 14:00:59 -0800 Received: from ranma.dkp.com (ranma.dkp.com [205.150.40.12]) by mail.dkp.com (Postfix) with ESMTP id 4E0D21AB32 for ; Wed, 5 Dec 2001 16:00:58 -0500 (EST) Received: by ranma.dkp.com (Postfix, from userid 168) id 15B855AA0B; Wed, 5 Dec 2001 16:00:58 -0500 (EST) Date: Wed, 5 Dec 2001 16:00:57 -0500 From: Andrew Klaassen To: "Linux XFS (E-mail)" Subject: Re: compilers Message-ID: <20011205160057.G3832@dkp.com> Mail-Followup-To: "Linux XFS (E-mail)" References: <23D04BDBA646D411BDDD00D0B774B539046029AE@sa-bwmail1.esr.hp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <23D04BDBA646D411BDDD00D0B774B539046029AE@sa-bwmail1.esr.hp.com> User-Agent: Mutt/1.3.23i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 03:36:57PM -0500, Christian, Chip wrote: > Just to add another data point... > > I recently started using the 2.4.9-12 kernel on a Dell 2450 with an > Intel e1000 gig-ethernet card running RedHat Linux 7.1. > > The SGI-built kernels work. > The RedHat-built kernels work. > When I built my own kernels, the module would load, > the card would get ifconfig'd up, etc. Couldn't pass traffic. What were you using for an e1000 driver? Does it work if you use the one directly from Intel? (Just checking...) Andrew Klaassen From owner-linux-xfs@oss.sgi.com Wed Dec 5 14:07:53 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5M7re00617 for linux-xfs-outgoing; Wed, 5 Dec 2001 14:07:53 -0800 Received: from atlrel6.hp.com (atlrel6.hp.com [156.153.255.205]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5M7no00594 for ; Wed, 5 Dec 2001 14:07:49 -0800 Received: from xatlrelay1.atl.hp.com (xatlrelay1.atl.hp.com [15.45.89.190]) by atlrel6.hp.com (Postfix) with ESMTP id F31681F91C; Wed, 5 Dec 2001 16:07:38 -0500 (EST) Received: from sa-bwmail1.esr.hp.com (sa-bwmail1.esr.hp.com [15.1.192.32]) by xatlrelay1.atl.hp.com (Postfix) with ESMTP id E09FF1F516; Wed, 5 Dec 2001 16:07:32 -0500 (EST) Received: by sa-bwmail1.esr.hp.com with Internet Mail Service (5.5.2653.19) id ; Wed, 5 Dec 2001 16:07:04 -0500 Message-ID: <23D04BDBA646D411BDDD00D0B774B539046029B2@sa-bwmail1.esr.hp.com> From: "Christian, Chip" To: "'Andrew Klaassen'" , "Linux XFS (E-mail)" Subject: RE: compilers Date: Wed, 5 Dec 2001 16:07:04 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Just the one located in the kernel-source rpm. It's now under drivers/addon/e1000. -----Original Message----- From: Andrew Klaassen [mailto:ak@dkp.com] Sent: Wednesday, December 05, 2001 16:01 To: Linux XFS (E-mail) Subject: Re: compilers On Wed, Dec 05, 2001 at 03:36:57PM -0500, Christian, Chip wrote: > Just to add another data point... > > I recently started using the 2.4.9-12 kernel on a Dell 2450 with an > Intel e1000 gig-ethernet card running RedHat Linux 7.1. > > The SGI-built kernels work. > The RedHat-built kernels work. > When I built my own kernels, the module would load, > the card would get ifconfig'd up, etc. Couldn't pass traffic. What were you using for an e1000 driver? Does it work if you use the one directly from Intel? (Just checking...) Andrew Klaassen From owner-linux-xfs@oss.sgi.com Wed Dec 5 14:40:40 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5MeeM01260 for linux-xfs-outgoing; Wed, 5 Dec 2001 14:40:40 -0800 Received: from anime.net (root@anime.net [63.172.78.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Mebo01234 for ; Wed, 5 Dec 2001 14:40:37 -0800 Received: from localhost (goemon@localhost) by anime.net (8.11.6/8.9.3) with ESMTP id fB5LeUi07858; Wed, 5 Dec 2001 13:40:30 -0800 Date: Wed, 5 Dec 2001 13:40:29 -0800 (PST) From: Dan Hollis To: Simon Matter cc: linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? In-Reply-To: <3C0E1B6C.F4518BDD@ch.sauter-bc.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 5 Dec 2001, Simon Matter wrote: > Dan Hollis schrieb: > > How can one xfs_check/xfs_repair a root filesystem? > Boot from a bootable CD and perform the repair on the unmounted > filesystem. > Have a look at this http://lbt.linuxcare.com/index.epl Seems a bit ridiculous to have to install a CDrom just to fsck the root filesystem. After all, reiserfs can reiserfsck a filesystem mounted read-only... Maybe it's a bug in xfs_check/xfs_repair? -Dan -- [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Wed Dec 5 14:50:41 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5Mofb01707 for linux-xfs-outgoing; Wed, 5 Dec 2001 14:50:41 -0800 Received: from anime.net (root@anime.net [63.172.78.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Moco01685 for ; Wed, 5 Dec 2001 14:50:38 -0800 Received: from localhost (goemon@localhost) by anime.net (8.11.6/8.9.3) with ESMTP id fB5Lnok08007; Wed, 5 Dec 2001 13:49:50 -0800 Date: Wed, 5 Dec 2001 13:49:50 -0800 (PST) From: Dan Hollis To: Harri Haataja cc: Xianglong Yuan , Paulo Sergio Lemes Queiroz , Subject: Re: Files on XFS not safe?! In-Reply-To: <20011205175647.A4895@cs.helsinki.fi> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 5 Dec 2001, Harri Haataja wrote: > On Wed, Dec 05, 2001 at 10:47:41AM -0500, Xianglong Yuan wrote: > > >-Paulo Sergio Lemes Queiroz [05 Dec 2001 13:15 -0200] wrote: > > > http://oss.sgi.com/projects/xfs/faq.html#nulls > > Thanks. Is this true for all the journaling FS, say ReiserFS, > > JFS? > Maybe I might bring up this > http://kerneltrap.com/article.php?sid=389 FWIW I never had reiserfs nulling files... it is either more resistant to such situations or it operates in a totally different way than XFS. -Dan -- [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Wed Dec 5 14:58:27 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5MwRh01960 for linux-xfs-outgoing; Wed, 5 Dec 2001 14:58:27 -0800 Received: from relay-4v.club-internet.fr (relay-4v.club-internet.fr [194.158.96.115]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5MwOo01935 for ; Wed, 5 Dec 2001 14:58:24 -0800 Received: from club-internet.fr (bas52-134.idf.w.club-internet.fr [212.194.36.134]) by relay-4v.club-internet.fr (Postfix) with ESMTP id C2E431687; Wed, 5 Dec 2001 22:58:15 +0100 (CET) Message-ID: <3C0E9977.7EEF4D8@club-internet.fr> Date: Wed, 05 Dec 2001 23:02:31 +0100 From: Jean Francois Martinez X-Mailer: Mozilla 4.78 [en] (X11; U; Linux 2.4.9-13SGI_XFS_PR1 i586) X-Accept-Language: en MIME-Version: 1.0 To: Austin Gonyou Cc: Igor Lautar , "'Christian, Chip'" , "Linux XFS (E-mail)" Subject: Re: compilers References: <1007585421.2026.0.camel@UberGeek> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Austin Gonyou wrote: > > This is not true. RH 7.1 gcc is not broke(as in can't compile code that > runs), but there was an update to it which fixed certain issues. You > definitely CAN compile kernels with that default gcc though that will > boot. Will you have problems? ymmv. The best course of action is to > either get gcc3, or update your current gcc from updates.redhat.com. > Gcc3 is known to NOT work with kernel.  -- Jean Francois Martinez Project Independence http://independence.seul.org Because Linux should be for everyone From owner-linux-xfs@oss.sgi.com Wed Dec 5 15:11:48 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5NBmK02377 for linux-xfs-outgoing; Wed, 5 Dec 2001 15:11:48 -0800 Received: from mitta.telekabel.at (212186140225.15.wu-wien.teleweb.at [212.186.140.225]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5NBho02355 for ; Wed, 5 Dec 2001 15:11:44 -0800 Received: from yahoo.com.au (localhost [127.0.0.1]) by mitta.telekabel.at (8.12.1/8.12.1) with ESMTP id fB5MBdJ1000556; Wed, 5 Dec 2001 23:11:40 +0100 Message-ID: <3C0E9B9B.80203@yahoo.com.au> Date: Wed, 05 Dec 2001 23:11:39 +0100 From: Adam Cioccarelli User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.6) Gecko/20011120 X-Accept-Language: en-au, en, en-gb, en-us, de MIME-Version: 1.0 To: Jean Francois Martinez CC: "Linux XFS (E-mail)" Subject: Re: compilers References: <1007585421.2026.0.camel@UberGeek> <3C0E9977.7EEF4D8@club-internet.fr> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk In what way? I have been using gcc 3.0 and gcc 3.0.2 for some time now without any problems... Adam Jean Francois Martinez wrote: > Austin Gonyou wrote: > >>This is not true. RH 7.1 gcc is not broke(as in can't compile code that >>runs), but there was an update to it which fixed certain issues. You >>definitely CAN compile kernels with that default gcc though that will >>boot. Will you have problems? ymmv. The best course of action is to >>either get gcc3, or update your current gcc from updates.redhat.com. >> >> > > Gcc3 is known to NOT work with kernel. >  > From owner-linux-xfs@oss.sgi.com Wed Dec 5 15:16:28 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5NGST03398 for linux-xfs-outgoing; Wed, 5 Dec 2001 15:16:28 -0800 Received: from anime.net (root@anime.net [63.172.78.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5NGPo03376 for ; Wed, 5 Dec 2001 15:16:25 -0800 Received: from localhost (goemon@localhost) by anime.net (8.11.6/8.9.3) with ESMTP id fB5MGMF08557; Wed, 5 Dec 2001 14:16:22 -0800 Date: Wed, 5 Dec 2001 14:16:22 -0800 (PST) From: Dan Hollis To: Xianglong Yuan cc: Subject: Re: Follow up -- Re: Files on XFS not safe?! In-Reply-To: <20011205140358.F3446@real-uma.mit.edu> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 5 Dec 2001, Xianglong Yuan wrote: > I just cross-by an excellent article on various journaling > techniques used in ext3. > http://www-106.ibm.com/developerworks/linux/library/l-fs8.html It looks like tux2 is taking a different approach, and one of the goals is total data integrity -- not just metadata. They are using phase trees. It's been described as "failsafe" whereas XFS and most other journaling fs are just "crash resistant". http://people.nl.linux.org/~phillips/tux2/phase.tree.tutorial.html -Dan -- [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Wed Dec 5 15:23:04 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5NN4U03663 for linux-xfs-outgoing; Wed, 5 Dec 2001 15:23:04 -0800 Received: from relay-4v.club-internet.fr (relay-4v.club-internet.fr [194.158.96.115]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5NMwo03633 for ; Wed, 5 Dec 2001 15:22:58 -0800 Received: from club-internet.fr (bas52-134.idf.w.club-internet.fr [212.194.36.134]) by relay-4v.club-internet.fr (Postfix) with ESMTP id 4E8401778; Wed, 5 Dec 2001 23:22:55 +0100 (CET) Message-ID: <3C0E9F40.89749EB6@club-internet.fr> Date: Wed, 05 Dec 2001 23:27:12 +0100 From: Jean Francois Martinez X-Mailer: Mozilla 4.78 [en] (X11; U; Linux 2.4.9-13SGI_XFS_PR1 i586) X-Accept-Language: en MIME-Version: 1.0 To: Adam Cioccarelli Cc: "Linux XFS (E-mail)" Subject: Re: compilers References: <1007585421.2026.0.camel@UberGeek> <3C0E9977.7EEF4D8@club-internet.fr> <3C0E9B9B.80203@yahoo.com.au> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Adam Cioccarelli wrote: > > In what way? > > I have been using gcc 3.0 and gcc 3.0.2 for some time now without any > problems... > That is until you compile/use the wrong module. Do an "rpm -ba kernel.spec" in order to ensure you compile everything or about evrything. (And hack the SPEC file for forcing use of gcc). In Linux's history there has been no case of a new compiler not breaking kernel. Not necessarily compiler's fault since there is non compliant code rejected by newer compilers (they also happen to be more compliant) and some cases where kernel code relies on compiler not being too smart. > Adam > > Jean Francois Martinez wrote: > > > Austin Gonyou wrote: > > > >>This is not true. RH 7.1 gcc is not broke(as in can't compile code that > >>runs), but there was an update to it which fixed certain issues. You > >>definitely CAN compile kernels with that default gcc though that will > >>boot. Will you have problems? ymmv. The best course of action is to > >>either get gcc3, or update your current gcc from updates.redhat.com. > >> > >> > > > > Gcc3 is known to NOT work with kernel. > >  > > -- Jean Francois Martinez Project Independence http://independence.seul.org Because Linux should be for everyone From owner-linux-xfs@oss.sgi.com Wed Dec 5 15:30:23 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5NUNX03938 for linux-xfs-outgoing; Wed, 5 Dec 2001 15:30:23 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5NUIo03915 for ; Wed, 5 Dec 2001 15:30:18 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id OAA07178 for ; Wed, 5 Dec 2001 14:30:19 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id QAA3775769; Wed, 5 Dec 2001 16:28:59 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id QAA33107; Wed, 5 Dec 2001 16:28:59 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB5MSaJ03068; Wed, 5 Dec 2001 16:28:36 -0600 Subject: Re: How to fsck read-only-mounted root filesystem? From: Steve Lord To: Dan Hollis Cc: Simon Matter , linux xfs ml In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 16:28:36 -0600 Message-Id: <1007591316.30179.10.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 2001-12-05 at 15:40, Dan Hollis wrote: > On Wed, 5 Dec 2001, Simon Matter wrote: > > Dan Hollis schrieb: > > > How can one xfs_check/xfs_repair a root filesystem? > > Boot from a bootable CD and perform the repair on the unmounted > > filesystem. > > Have a look at this http://lbt.linuxcare.com/index.epl > > Seems a bit ridiculous to have to install a CDrom just to fsck the root > filesystem. No, it is just that the environment xfs comes from (Irix) tends to use other mechanisms to fix root filesystems - like an alternate root, or a miniroot. It is also not 'install a cdrom', it is boot with a bootable cdrom. Doing a user space fixup of a live filesystem also relies upon the block device cache and the filesystem's cache of metadata being coherent. If they are not then this approach does not work and can lead to nasty crashes very quickly. XFS was until recently using completely different caches for metadata than the block interface, they now use the same memory - mostly. Until they are coherent there is absolutely no point in doing modifications to a mounted XFS fs from user space. > > After all, reiserfs can reiserfsck a filesystem mounted read-only... > > Maybe it's a bug in xfs_check/xfs_repair? It is deliberate. Steve > > -Dan > -- > [-] Omae no subete no kichi wa ore no mono da. [-] -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 15:40:39 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5Nedx04376 for linux-xfs-outgoing; Wed, 5 Dec 2001 15:40:39 -0800 Received: from fep02-app.kolumbus.fi (fep02-0.kolumbus.fi [193.229.0.44]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5NeZo04354 for ; Wed, 5 Dec 2001 15:40:35 -0800 Received: from there ([62.248.183.118]) by fep02-app.kolumbus.fi (InterMail vM.5.01.03.08 201-253-122-118-108-20010628) with SMTP id <20011205224032.VRAN4415.fep02-app.kolumbus.fi@there>; Thu, 6 Dec 2001 00:40:32 +0200 Content-Type: text/plain; charset="iso-8859-1" From: Hristo Grigorov To: Jean Francois Martinez , Austin Gonyou Subject: Re: compilers Date: Thu, 6 Dec 2001 00:42:19 +0200 X-Mailer: KMail [version 1.3.1] Cc: Igor Lautar , "'Christian, Chip'" , "Linux XFS (E-mail)" References: <1007585421.2026.0.camel@UberGeek> <3C0E9977.7EEF4D8@club-internet.fr> In-Reply-To: <3C0E9977.7EEF4D8@club-internet.fr> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <20011205224032.VRAN4415.fep02-app.kolumbus.fi@there> Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thursday 06 December 2001 00:02, Jean Francois Martinez wrote: > Austin Gonyou wrote: > > This is not true. RH 7.1 gcc is not broke(as in can't compile code that > > runs), but there was an update to it which fixed certain issues. You > > definitely CAN compile kernels with that default gcc though that will > > boot. Will you have problems? ymmv. The best course of action is to > > either get gcc3, or update your current gcc from updates.redhat.com. > > Gcc3 is known to NOT work with kernel. Linux version 2.4.16-xfs (root@magdanoz) (gcc version 3.0.2 20011002 (Red Hat Linux 7.1 3.0.1-4)) Really ? How does it happen that it works here for months already ? Not even one single fault. And the system is not idle for sure.... -- Cheers, Hristo. From owner-linux-xfs@oss.sgi.com Wed Dec 5 15:50:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB5No3k04599 for linux-xfs-outgoing; Wed, 5 Dec 2001 15:50:03 -0800 Received: from mailhost.idcomm.com (mailhost.idcomm.com [207.40.196.14]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB5Nnxo04571 for ; Wed, 5 Dec 2001 15:49:59 -0800 Received: from idcomm.com (IDENT:SsmaMiMu+GOYZ/UMQIG2qgij9qGPWkNt@k56-pip12.idcomm.com [209.60.72.139]) by mailhost.idcomm.com (8.10.2/8.10.0) with ESMTP id fB5MpQE22896 for ; Wed, 5 Dec 2001 15:51:26 -0700 Message-ID: <3C0EA4DB.82F7FB0B@idcomm.com> Date: Wed, 05 Dec 2001 15:51:07 -0700 From: "D. Stimits" Reply-To: stimits@idcomm.com X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.6-pre1-xfs-4 i686) X-Accept-Language: en MIME-Version: 1.0 To: "XFS: linux-xfs@oss.sgi.com" Subject: Re: compilers References: <1007585421.2026.0.camel@UberGeek> <3C0E9977.7EEF4D8@club-internet.fr> <20011205224032.VRAN4415.fep02-app.kolumbus.fi@there> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hristo Grigorov wrote: > > On Thursday 06 December 2001 00:02, Jean Francois Martinez wrote: > > Austin Gonyou wrote: > > > This is not true. RH 7.1 gcc is not broke(as in can't compile code that > > > runs), but there was an update to it which fixed certain issues. You > > > definitely CAN compile kernels with that default gcc though that will > > > boot. Will you have problems? ymmv. The best course of action is to > > > either get gcc3, or update your current gcc from updates.redhat.com. > > > > Gcc3 is known to NOT work with kernel. > > Linux version 2.4.16-xfs (root@magdanoz) (gcc version 3.0.2 20011002 (Red Hat Linux 7.1 3.0.1-4)) > > Really ? How does it happen that it works here for months already ? Not even > one single fault. And the system is not idle for sure.... > > -- > Cheers, Hristo. You can't say it works unless you have compiled and tested all configurations with all drivers. All you can say is you got lucky with your particular configuration. As soon as you use the wrong hardware or config, it won't work anymore. D. Stimits, stimits@idcomm.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 16:11:14 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB60BEk05661 for linux-xfs-outgoing; Wed, 5 Dec 2001 16:11:14 -0800 Received: from mailhost.idcomm.com (mailhost.idcomm.com [207.40.196.14]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB60BAo05639 for ; Wed, 5 Dec 2001 16:11:10 -0800 Received: from idcomm.com (IDENT:RamFkMn/L/qBC1OZCAC6mUlOvZvoy2US@k56-pip25.idcomm.com [209.60.72.152]) by mailhost.idcomm.com (8.10.2/8.10.0) with ESMTP id fB5NCcE28029 for ; Wed, 5 Dec 2001 16:12:38 -0700 Message-ID: <3C0EA9D1.53E1E1E8@idcomm.com> Date: Wed, 05 Dec 2001 16:12:17 -0700 From: "D. Stimits" Reply-To: stimits@idcomm.com X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.6-pre1-xfs-4 i686) X-Accept-Language: en MIME-Version: 1.0 CC: linux-xfs@oss.sgi.com Subject: Re: Follow up -- Re: Files on XFS not safe?! References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Dan Hollis wrote: > > On Wed, 5 Dec 2001, Xianglong Yuan wrote: > > I just cross-by an excellent article on various journaling > > techniques used in ext3. > > http://www-106.ibm.com/developerworks/linux/library/l-fs8.html > > It looks like tux2 is taking a different approach, and one of the goals is > total data integrity -- not just metadata. They are using phase trees. > It's been described as "failsafe" whereas XFS and most other journaling fs > are just "crash resistant". > > http://people.nl.linux.org/~phillips/tux2/phase.tree.tutorial.html > > -Dan > -- > [-] Omae no subete no kichi wa ore no mono da. [-] I'm curious what kind of journaling is used with NTFS, if anyone knows? I assume it is probably meta. D. Stimits, stimits@idcomm.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 16:30:13 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB60UDK06982 for linux-xfs-outgoing; Wed, 5 Dec 2001 16:30:13 -0800 Received: from chef.cc.absoval.com (cpe-66-1-218-101.fl.sprintbbd.net [66.1.218.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB60U7o06960 for ; Wed, 5 Dec 2001 16:30:08 -0800 Received: from ieee.org (IDENT:bs@thebs.cc.absoval.com [192.168.100.89]) by chef.cc.absoval.com (8.9.3/8.9.3) with ESMTP id SAA23324; Wed, 5 Dec 2001 18:29:14 -0500 Message-ID: <3C0EADE2.8028F54E@ieee.org> Date: Wed, 05 Dec 2001 18:29:38 -0500 From: Bryan-TheBS-Smith Organization: SmithConcepts/AbsoluteValueSystems X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.2.19-6.2.12 i686) X-Accept-Language: en MIME-Version: 1.0 To: stimits@idcomm.com CC: linux-xfs@oss.sgi.com Subject: Re: Follow up -- Re: Files on XFS not safe?! References: <3C0EA9D1.53E1E1E8@idcomm.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk "D. Stimits" wrote: > I'm curious what kind of journaling is used with NTFS, if anyone knows? > I assume it is probably meta. It's meta, and "overly aggressive" in recovery. I.e., I've had two production NT 4.0 servers go to the journal when they should have done a full chkdsk instead. Both times the filesystems were toasted. Those were the last two NT servers I ever installed. NTFS is based on OS/2's HPFS from IBM. And from what I've heard (*DISCLAIMER*: I could be spreading unintentional FUD here), IBM's JFS is more of the same. I guess it is just a design consideration. I've been running Ext3 since early 2000 on over 100 systems, and XFS since February of this year on over a dozen. What I've always liked about Ext3 is that it's not shy to go to an fsck. Since I haven't used XFS long enough, nor completely understand its approach, I cannot comment. But I haven't had an issue with XFS yet either. -- TheBS -- Bryan "TheBS" Smith mailto:b.j.smith@ieee.org chat:thebs413 Engineer AbsoluteValue Systems, Inc. http://www.linux-wlan.org President SmithConcepts, Inc. http://www.SmithConcepts.com ---------------------------------------------------------------- "The [US] Constitution guarantees you Free, not Fair. 'Fair' is a socialist concept." -- Shawn McMahon From owner-linux-xfs@oss.sgi.com Wed Dec 5 16:52:06 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB60q6b07361 for linux-xfs-outgoing; Wed, 5 Dec 2001 16:52:06 -0800 Received: from k-7.stesmi.com (IDENT:root@as4-1-7.has.s.bonet.se [217.215.31.238]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB60pvo07339 for ; Wed, 5 Dec 2001 16:51:57 -0800 Received: from stesmi.com (voyager.stesmi.com [192.168.1.11]) by k-7.stesmi.com (8.11.2/8.8.7) with ESMTP id fB5NoiG29813; Thu, 6 Dec 2001 00:50:44 +0100 Message-ID: <3C0EB34C.4010904@stesmi.com> Date: Thu, 06 Dec 2001 00:52:44 +0100 From: Stefan Smietanowski User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.6) Gecko/20011120 X-Accept-Language: en-us MIME-Version: 1.0 To: Hristo Grigorov CC: Jean Francois Martinez , Austin Gonyou , Igor Lautar , "'Christian, Chip'" , "Linux XFS (E-mail)" Subject: Re: compilers References: <1007585421.2026.0.camel@UberGeek> <3C0E9977.7EEF4D8@club-internet.fr> <20011205224032.VRAN4415.fep02-app.kolumbus.fi@there> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi. >>>This is not true. RH 7.1 gcc is not broke(as in can't compile code that >>>runs), but there was an update to it which fixed certain issues. You >>>definitely CAN compile kernels with that default gcc though that will >>>boot. Will you have problems? ymmv. The best course of action is to >>>either get gcc3, or update your current gcc from updates.redhat.com. >>> >>Gcc3 is known to NOT work with kernel. >> > > Linux version 2.4.16-xfs (root@magdanoz) (gcc version 3.0.2 20011002 (Red Hat Linux 7.1 3.0.1-4)) > > Really ? How does it happen that it works here for months already ? Not even > one single fault. And the system is not idle for sure.... It works for you, have you tried installing every single driver in the kernel tree? You 100% sure you trust all drivers? I'm not talking about broken code, I'm talking compiler problems. Also, how can you use 3.0.2 for months already? Was it the ISDN subsystem that didn't work with gcc3? Something didn't work as it should. Didn't even compile but the compiler itself died with an internal compiler fault. And lately there has been kernel modifications to make use of gcc. XFS for one has had a fix. (Tell me if I'm wrong here). I sometimes use gcc3 to compile the kernel also, but that doesn't mean I trust it on a mission-critical server. // Stefan From owner-linux-xfs@oss.sgi.com Wed Dec 5 16:55:45 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB60tjt07524 for linux-xfs-outgoing; Wed, 5 Dec 2001 16:55:45 -0800 Received: from anime.net (root@anime.net [63.172.78.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB60tfo07502 for ; Wed, 5 Dec 2001 16:55:41 -0800 Received: from localhost (goemon@localhost) by anime.net (8.11.6/8.9.3) with ESMTP id fB5NtW910313; Wed, 5 Dec 2001 15:55:32 -0800 Date: Wed, 5 Dec 2001 15:55:32 -0800 (PST) From: Dan Hollis To: Steve Lord cc: Simon Matter , linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? In-Reply-To: <1007591316.30179.10.camel@jen.americas.sgi.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On 5 Dec 2001, Steve Lord wrote: > It is also not 'install a cdrom', it is boot with a bootable cdrom. This machine doesnt have any cdrom drive, so it is "install a cdrom drive" just to fix the root fs. :-( > > After all, reiserfs can reiserfsck a filesystem mounted read-only... > > Maybe it's a bug in xfs_check/xfs_repair? > It is deliberate. So is changing this behaviour in the "TODO" list, or is it to permanently remain so? It was *really* suprising to me that you can't repair a read-only mounted filesystem, I suspect it will be equally suprising to others. -Dan -- [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Wed Dec 5 17:29:53 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB61TrR11307 for linux-xfs-outgoing; Wed, 5 Dec 2001 17:29:53 -0800 Received: from zok.sgi.com (zok.SGI.COM [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB61Tmo11285 for ; Wed, 5 Dec 2001 17:29:48 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with SMTP id fB60TdA29758 for ; Wed, 5 Dec 2001 16:29:39 -0800 Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id LAA12861; Thu, 6 Dec 2001 11:28:22 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id LAA50151; Thu, 6 Dec 2001 11:28:20 +1100 (AEDT) Date: Thu, 6 Dec 2001 11:28:20 +1100 From: Nathan Scott To: David Wasylciw Cc: linux-xfs@oss.sgi.com Subject: Re: Problems setting ACL with setfacl Message-ID: <20011206112820.B50483@wobbly.melbourne.sgi.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from david@wasylciw.com on Wed, Dec 05, 2001 at 01:50:51AM -0400 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi, On Wed, Dec 05, 2001 at 01:50:51AM -0400, David Wasylciw wrote: > Hello All, > I've recently installed kernel 2.4.16 with the latest XFS all together patch > and of course compiled in ACL support. I'm running Redhat 6.2 and have > installed all the latest acl programs and libraries but for some reason when > I attempt to add a user to the ACL list for a file I get the following > error: > setfacl: file: Resulting ACL > `,user::rw-,user:dwasyl:rw-,mask::---,other::r--': Invalid entry type at > entry 1 This error message is pretty unhelpful - it will get better in the future (when we have a different libacl). > This was generated with the command: setfacl -m u:dwasyl:rw testfile > I checked through the archives and saw something about perhaps needing to > have the permissions listed as rw- but that did not seem to help and > generated the exact same error. > > It would be much appreciated if anyone could help point me in the right > direction and if any more information is needed just let me know. You probably don't have a mask entry in the ACL. Try something like this: $ setfacl -m m::rwx testfile $ setfacl -m u:dwasyl:rw- testfile The above recipe works for me. Andreas' man pages are quite extensive and have a section describing the mask entry - see: http://acl.bestbits.at/man/acl.5.html cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Wed Dec 5 17:46:07 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB61k7511764 for linux-xfs-outgoing; Wed, 5 Dec 2001 17:46:07 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB61k4o11742 for ; Wed, 5 Dec 2001 17:46:04 -0800 Received: from sherman.melbourne.sgi.com (sherman.melbourne.sgi.com [134.14.55.175]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id QAA03185 for ; Wed, 5 Dec 2001 16:46:01 -0800 (PST) mail_from (kaos@sherman.melbourne.sgi.com) Received: (from kaos@localhost) by sherman.melbourne.sgi.com (8.9.3/8.9.3) id LAA31898; Thu, 6 Dec 2001 11:45:59 +1100 Date: Thu, 6 Dec 2001 11:45:59 +1100 From: Keith Owens Message-Id: <200112060045.LAA31898@sherman.melbourne.sgi.com> Subject: TAKE - Remove misc patch that is now supplied by ia64 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk The ia64 patch now takes care of exporting irq_stat. Date: Wed Dec 5 16:44:28 PST 2001 Workarea: sherman.melbourne.sgi.com:/build/kaos/2.4.x-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107901a linux/kernel/ksyms.c - 1.121 linux/fs/devfs/base.c - 1.24 From owner-linux-xfs@oss.sgi.com Wed Dec 5 17:48:22 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB61mMT11920 for linux-xfs-outgoing; Wed, 5 Dec 2001 17:48:22 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB61mHo11898 for ; Wed, 5 Dec 2001 17:48:17 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via SMTP id QAA12590 for ; Wed, 5 Dec 2001 16:48:03 -0800 (PST) mail_from (ivanr@sgi.com) From: ivanr@sgi.com Received: from omen.melbourne.sgi.com (omen.melbourne.sgi.com [134.14.55.139]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id LAA12963; Thu, 6 Dec 2001 11:46:57 +1100 Received: from localhost (ivanr@localhost) by omen.melbourne.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id LAA65463; Thu, 6 Dec 2001 11:46:57 +1100 (EST) X-Authentication-Warning: omen.melbourne.sgi.com: ivanr owned process doing -bs Date: Thu, 6 Dec 2001 11:46:56 +1100 X-X-Sender: ivanr@omen.melbourne.sgi.com To: Matthijs van der Klip cc: Linux XFS Mailing List Subject: Re: using xfsdump to synchronise filesystems? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 5 Dec 2001, Matthijs van der Klip wrote: > I know this is a bit of a crazy idea, but I'd like to know if something > like this is (remotely) possible... > > 1) Take a server with a 10Gb (max) volume. This volume is live: files are > being added/deleted through ftp (Read/Write). > > 2) Take x clients which need to receive a synchronised copy of the > mentioned volume above. Contents have to be served out through http > (Read only). > > Would it be possible to say make a level 0 dump every night, a level 1 > dump every hour, and a level 2 dump every 5 minutes (or use even more > levels?) and then restore these dumps on the clients to keep their volumes > synchronised? > > Reason I'm asking this is I'm looking for a way to distribute only changes > in a filesystem. Rsync has to scan the whole filesystem and just takes to > long. I don't think xfsdump is a very good choice here. xfsdump will scan the entire filesystem every time you run it, even if you select a higher level dump. I doubt that for a 10GB filesystem it will be able to run to completion within 5 minutes ('course this depends on your circumstances). Try timing this: xfsdump -L session -M media -f - /path/to/filesystem > /dev/null and see how long it takes. Specifying higher levels for xfsdump only reduces the amount of data that it writes out, it will not necessarily reduce it's running time. Certainly the initial setup phase will take longer (how long depends on the filesystem). I think you'd be better off monitoring the ftp log to see what files have changed and then use rsync on those. I'm sure it'd be fairly easy to do this in something like perl. Ivan -- Ivan Rayner ivanr@sgi.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 18:08:57 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB628vI12476 for linux-xfs-outgoing; Wed, 5 Dec 2001 18:08:57 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB628ro12454 for ; Wed, 5 Dec 2001 18:08:53 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via SMTP id RAA07529 for ; Wed, 5 Dec 2001 17:08:54 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id MAA13105; Thu, 6 Dec 2001 12:07:34 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id MAA50945; Thu, 6 Dec 2001 12:07:33 +1100 (AEDT) Date: Thu, 6 Dec 2001 12:07:33 +1100 From: Nathan Scott To: Steve Lord , "Quang Nguyen (Ngo)" Cc: linux-xfs@oss.sgi.com Subject: Re: Extended Attributes Message-ID: <20011206120732.C50483@wobbly.melbourne.sgi.com> References: <1007577855.22718.24.camel@jen.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <1007577855.22718.24.camel@jen.americas.sgi.com>; from lord@sgi.com on Wed, Dec 05, 2001 at 12:44:15PM -0600 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 12:44:15PM -0600, Steve Lord wrote: > On Wed, 2001-12-05 at 12:39, Quang Nguyen (Ngo) wrote: > > > > Does anyone know if other journaling file systems also support extended > > attributes besides XFS? > > Given there is a patch for ext2, it is not inconceivble that ext3 > will support them at some point. There's a patch for ext3 now, based on an -ac tree, from the same folk who do the ext2 version (http://acl.bestbits.at/). > Not sure about jfs - it may on os2, but probably does not on Linux. The IBM guys tell me there is code in their tree to support extended attributes and ACLs, but my understanding is that at this stage it is not compiled in. There would be some work to get it to tie in with whatever interface eventually gets put into the kernel of course... I guess they're waiting to see how that pans out before working on that area. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Wed Dec 5 18:15:41 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB62Ffv12731 for linux-xfs-outgoing; Wed, 5 Dec 2001 18:15:41 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB62FKo12707 for ; Wed, 5 Dec 2001 18:15:21 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via SMTP id CAA1399119 for ; Thu, 6 Dec 2001 02:15:02 +0100 (CET) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id MAA13146; Thu, 6 Dec 2001 12:13:35 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id MAA23845; Thu, 6 Dec 2001 12:13:35 +1100 (AEDT) Date: Thu, 6 Dec 2001 12:13:34 +1100 From: Nathan Scott To: Juer Lee Cc: linux-xfs@oss.sgi.com Subject: Re: Syscall number Message-ID: <20011206121334.D50483@wobbly.melbourne.sgi.com> References: <1007577444.28030.7.camel@stout.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <1007577444.28030.7.camel@stout.americas.sgi.com>; from sandeen@sgi.com on Wed, Dec 05, 2001 at 12:37:24PM -0600 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 12:37:24PM -0600, Eric Sandeen wrote: > Hi Juer - > > There is an ongoing discussion on LKML about this very subject - some of > it has been cross-posted to this list as well. > > The short answer is that there are not standard numbers available yet, > but it's in the works. > > -Eric > > On Mon, 2001-12-03 at 06:03, Juer Lee wrote: > > Hi, All ( XFS team ), > > > > We have been testing XFS ACL on Power PC with linux-2.4.5-xfs-1.0.1 for > > 4 months already, we still haven't found any big problems, so it seems > > that it is the time to add syscall number for ACL. > > There are three syscall number which need to be added: __NR__attrctl, > > __NR__acl_get and __NR__acl_set. > > The one thing we do know is that there will need to be an interface change, so it wont be these three syscalls that we end up with. > > > > Can anybody tell me how to apply them as standard numbers? > > Nope - be sure to let us know if you find out how! ;) cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Wed Dec 5 18:31:59 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB62VxG13108 for linux-xfs-outgoing; Wed, 5 Dec 2001 18:31:59 -0800 Received: from rj.sgi.com (rj.sgi.com [204.94.215.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB62Vuo13086 for ; Wed, 5 Dec 2001 18:31:56 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by rj.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with SMTP id fB61VnY30654 for ; Wed, 5 Dec 2001 17:31:49 -0800 Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id MAA13263; Thu, 6 Dec 2001 12:30:33 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id MAA39411; Thu, 6 Dec 2001 12:30:32 +1100 (AEDT) Date: Thu, 6 Dec 2001 12:30:31 +1100 From: Nathan Scott To: Austin Gonyou Cc: "Quang Nguyen (Ngo)" , "'linux-xfs@oss.sgi.com'" Subject: Re: XFS ACL implementation. Message-ID: <20011206123031.E50483@wobbly.melbourne.sgi.com> References: <1007583963.3687.4.camel@UberGeek> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <1007583963.3687.4.camel@UberGeek>; from austin@coremetrics.com on Wed, Dec 05, 2001 at 02:26:03PM -0600 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi, On Wed, Dec 05, 2001 at 02:26:03PM -0600, Austin Gonyou wrote: > I have a question on the ACL implementation in XFS, is it possible to [ie. you have a question on POSIX ACLs] > have "roles" setup so certain people can access certain filesystems, > etc, regardless of the higher group perms? If I understand what you're asking, then I think so, yes. There's an example ACL usage scenario which I think maps to what you mean by "roles" over here: http://acl.bestbits.at/example.html Hope this helps. The people on the acl-devel list will likely be able to give you better insight into applying ACLs in everyday use. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Wed Dec 5 19:05:05 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6355D13864 for linux-xfs-outgoing; Wed, 5 Dec 2001 19:05:05 -0800 Received: from UberGeek ([209.184.141.163]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB634vo13840 for ; Wed, 5 Dec 2001 19:04:57 -0800 Received: (qmail 3104 invoked by uid 500); 6 Dec 2001 02:03:23 -0000 Subject: Re: Follow up -- Re: Files on XFS not safe?! From: Austin Gonyou To: stimits@idcomm.com Cc: linux-xfs@oss.sgi.com In-Reply-To: <3C0EA9D1.53E1E1E8@idcomm.com> References: <3C0EA9D1.53E1E1E8@idcomm.com> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-22sAfjMh+4n9v+uTYmj8" X-Mailer: Evolution/1.0 (Preview Release) Date: 05 Dec 2001 20:03:23 -0600 Message-Id: <1007604203.2985.0.camel@UberGeek> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk --=-22sAfjMh+4n9v+uTYmj8 Content-Type: text/plain Content-Transfer-Encoding: quoted-printable It is wholly possible that NTFS could be like OpenBSD's filesystem, in that it is technically not "journaling", but is soft updates? This would offer the stability, but at a speed hit, which seems to be right on part with Windows(tm) usual stuff.=20=20 On Wed, 2001-12-05 at 17:12, D. Stimits wrote: > Dan Hollis wrote: > >=20 > > On Wed, 5 Dec 2001, Xianglong Yuan wrote: > > > I just cross-by an excellent article on various journaling > > > techniques used in ext3. > > > http://www-106.ibm.com/developerworks/linux/library/l-fs8.html > >=20 > > It looks like tux2 is taking a different approach, and one of the > goals is > > total data integrity -- not just metadata. They are using phase trees. > > It's been described as "failsafe" whereas XFS and most other > journaling fs > > are just "crash resistant". > >=20 > > http://people.nl.linux.org/~phillips/tux2/phase.tree.tutorial.html > >=20 > > -Dan > > -- > > [-] Omae no subete no kichi wa ore no mono da. [-] >=20 > I'm curious what kind of journaling is used with NTFS, if anyone knows? > I assume it is probably meta. >=20 > D. Stimits, stimits@idcomm.com --=20 Austin Gonyou Systems Architect, CCNA=20 Coremetrics, Inc. Phone: 512-698-7250 email: austin@coremetrics.com =20 "Have regard for your name, since it will remain for you longer than a great store of gold." Ecclesiastes, Aprocrypha --=-22sAfjMh+4n9v+uTYmj8 Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQA8DtHr94g6ZVmFMoIRAt35AJ455h7TzywHnRQaI/kqzPMoVg/V9ACg2dj8 L5X4kDZ+XYlIsPGSNu7oTkc= =7zbB -----END PGP SIGNATURE----- --=-22sAfjMh+4n9v+uTYmj8-- From owner-linux-xfs@oss.sgi.com Wed Dec 5 19:40:22 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB63eMf14458 for linux-xfs-outgoing; Wed, 5 Dec 2001 19:40:22 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB63eHo14436 for ; Wed, 5 Dec 2001 19:40:18 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id SAA03786 for ; Wed, 5 Dec 2001 18:40:18 -0800 (PST) mail_from (lord@sgi.com) Received: from tulip-e185.americas.sgi.com (tulip-e185.americas.sgi.com [128.162.185.208]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id UAA3780441; Wed, 5 Dec 2001 20:38:59 -0600 (CST) Received: from sgi.com (eOBJAl+Km4RXRwLYh1MDAabsyWjDFldM@cf-vpn-sw-corp-64-23.corp.sgi.com [134.15.64.23]) by tulip-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id UAA11906; Wed, 5 Dec 2001 20:38:58 -0600 (CST) Message-ID: <3C0EDAA0.40208@sgi.com> Date: Wed, 05 Dec 2001 20:40:32 -0600 From: Stephen Lord User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.6) Gecko/20011120 X-Accept-Language: en-us MIME-Version: 1.0 To: Austin Gonyou CC: stimits@idcomm.com, linux-xfs@oss.sgi.com Subject: Re: Follow up -- Re: Files on XFS not safe?! References: <3C0EA9D1.53E1E1E8@idcomm.com> <1007604203.2985.0.camel@UberGeek> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Austin Gonyou wrote: >It is wholly possible that NTFS could be like OpenBSD's filesystem, in >that it is technically not "journaling", but is soft updates? This would >offer the stability, but at a speed hit, which seems to be right on part >with Windows(tm) usual stuff. > NTFS does journalling, but quite different from XFS - they record the state of a data structure before and after a metadata change - which makes it possible to unwind transactions. There is a book on it, but I am not sure it is still in print. Steve From owner-linux-xfs@oss.sgi.com Wed Dec 5 20:04:25 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB644PV14997 for linux-xfs-outgoing; Wed, 5 Dec 2001 20:04:25 -0800 Received: from starship.berlin (dsl-213-023-038-088.arcor-ip.net [213.23.38.88]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB644Lo14975 for ; Wed, 5 Dec 2001 20:04:21 -0800 Received: from daniel by starship.berlin with local (Exim 3.32 #1 (Debian)) id 16Boqr-0000m9-00; Thu, 06 Dec 2001 04:05:33 +0100 Content-Type: text/plain; charset="iso-8859-1" From: Daniel Phillips To: Nathan Scott , Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher Subject: Re: [PATCH] Revised extended attributes interface Date: Thu, 6 Dec 2001 04:05:32 +0100 X-Mailer: KMail [version 1.3.2] Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com References: <20011205143209.C44610@wobbly.melbourne.sgi.com> In-Reply-To: <20011205143209.C44610@wobbly.melbourne.sgi.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On December 5, 2001 04:32 am, Nathan Scott wrote: > Here is the revised interface. I believe it takes into account > the issues raised so far - further suggestions are also welcome, > of course. Hi Nathan, I still don't like the class parsing inside the kernel, it's hard to see what is good about that. Is there a difference between these two?: long sys_setxattr(char *path, char *name, void *value, size_t size, int flags) long sys_lsetxattr(char *path, char *name, void *value, size_t size, int flags) -- Daniel From owner-linux-xfs@oss.sgi.com Wed Dec 5 20:46:42 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB64kg715589 for linux-xfs-outgoing; Wed, 5 Dec 2001 20:46:42 -0800 Received: from mailhost.idcomm.com (mailhost.idcomm.com [207.40.196.14]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB64kco15567 for ; Wed, 5 Dec 2001 20:46:38 -0800 Received: from idcomm.com (IDENT:/HF3EJk5lIzGzJSWldbEXodOjnLq2zxe@k56-pip40.idcomm.com [209.60.72.167]) by mailhost.idcomm.com (8.10.2/8.10.0) with ESMTP id fB63m6E08036 for ; Wed, 5 Dec 2001 20:48:07 -0700 Message-ID: <3C0EEA62.CFD23105@idcomm.com> Date: Wed, 05 Dec 2001 20:47:46 -0700 From: "D. Stimits" Reply-To: stimits@idcomm.com X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.6-pre1-xfs-4 i686) X-Accept-Language: en MIME-Version: 1.0 CC: linux-xfs@oss.sgi.com Subject: Re: Follow up -- Re: Files on XFS not safe?! References: <3C0EA9D1.53E1E1E8@idcomm.com> <1007604203.2985.0.camel@UberGeek> <3C0EDAA0.40208@sgi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Stephen Lord wrote: > > Austin Gonyou wrote: > > >It is wholly possible that NTFS could be like OpenBSD's filesystem, in > >that it is technically not "journaling", but is soft updates? This would > >offer the stability, but at a speed hit, which seems to be right on part > >with Windows(tm) usual stuff. > > > > NTFS does journalling, but quite different from XFS - they record the > state of > a data structure before and after a metadata change - which makes it > possible > to unwind transactions. There is a book on it, but I am not sure it is > still in > print. > > Steve That might make an interesting "poor man's undelete". Since you couldn't stop other data from stepping on the old data location. However, it would be helpful to be able to trace steps back and undelete if data was still there...especially useful to know if it has not been altered so you can trust an undelete. D. Stimits, stimits@idcomm.com From owner-linux-xfs@oss.sgi.com Wed Dec 5 21:35:42 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB65ZgW16302 for linux-xfs-outgoing; Wed, 5 Dec 2001 21:35:42 -0800 Received: from picasso.visualfx.org ([24.100.161.53]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB65ZXo16280 for ; Wed, 5 Dec 2001 21:35:33 -0800 Received: from acm.org (localhost [127.0.0.1]) by picasso.visualfx.org (8.11.2/8.11.2) with ESMTP id fB64dJd30956; Wed, 5 Dec 2001 23:39:20 -0500 Message-ID: <3C0EF677.6B46487F@acm.org> Date: Wed, 05 Dec 2001 23:39:19 -0500 From: Andrew Ho X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.16-xfs i686) X-Accept-Language: en, zh MIME-Version: 1.0 To: Dean Roehrich CC: linux-xfs@oss.sgi.com Subject: xfsdump error Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Please find the following error. **************** /usr/src/linux-2.4-xfs/cmd/xfsdump 119 : make === include === make[1]: Nothing to be done for `default'. === librmt === make[1]: Nothing to be done for `default'. === common === make[1]: Nothing to be done for `default'. === copy === make[1]: Nothing to be done for `default'. === estimate === make[1]: Nothing to be done for `default'. === fsr === make[1]: Nothing to be done for `default'. === inventory === make[1]: Nothing to be done for `default'. === invutil === make[1]: Nothing to be done for `default'. === quota === make[1]: Nothing to be done for `default'. === dump === gcc -O1 -g -DDEBUG -funsigned-char -Wall -DDUMP -DRMT -DBASED -DDOSOCKS -DINVCONVFIX -DSIZEEST -DPIPEINVFIX -DEXTATTR -TATTR -I/usr/include/xfs -I/usr/include/attr '-DVERSION="1.1.9"' -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE -DXFS_BIG_FILES=1S_BIG_FILESYSTEMS=1 -I/usr/include/xfs -I/usr/include/attr -c -o attr.o attr.c attr.c: In function `attr_multi_by_handle': attr.c:57: `attr_op_t' undeclared (first use in this function) attr.c:57: (Each undeclared identifier is reported only once attr.c:57: for each function it appears in.) attr.c:57: `ops' undeclared (first use in this function) attr.c:57: warning: statement with no effect attr.c: In function `attr_list_by_handle': attr.c:121: `attr_op_t' undeclared (first use in this function) attr.c:121: parse error before "op" attr.c:126: `op' undeclared (first use in this function) attr.c:127: `ATTR_OP_IRIX_LIST' undeclared (first use in this function) make[1]: *** [attr.o] Error 1 make: *** [default] Error ************* I installed xfsprog package before I compile xfsdump. Andrew Ho From owner-linux-xfs@oss.sgi.com Wed Dec 5 22:43:14 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB66hEk17275 for linux-xfs-outgoing; Wed, 5 Dec 2001 22:43:14 -0800 Received: from zok.sgi.com (zok.SGI.COM [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB66h8o17246 for ; Wed, 5 Dec 2001 22:43:08 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with SMTP id fB65gwA06786 for ; Wed, 5 Dec 2001 21:42:58 -0800 Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id QAA14805; Thu, 6 Dec 2001 16:41:35 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id QAA51505; Thu, 6 Dec 2001 16:41:33 +1100 (AEDT) Date: Thu, 6 Dec 2001 16:41:31 +1100 From: Nathan Scott To: Daniel Phillips , Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [PATCH] Revised extended attributes interface Message-ID: <20011206164131.F50483@wobbly.melbourne.sgi.com> References: <20011205143209.C44610@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from phillips@bonn-fries.net on Thu, Dec 06, 2001 at 04:05:32AM +0100 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, Dec 06, 2001 at 04:05:32AM +0100, Daniel Phillips wrote: > Hi Nathan, > hey there. > I still don't like the class parsing inside the kernel, it's hard to see > what is good about that. I guess it ultimately comes down to simplicity. The IRIX interfaces have this separation of name and namespace - each operation has to indicate which namespace is to be used. That becomes very messy when you wish to work with multiple attribute names and namespaces at once. Since the namespace is intimately tied to the name anyway, this idea of specifying the two components together provides very clean APIs. The term "parsing" is a bit of an overstatement too. We're talking strncmp() complexity here, not lex/yacc. ;) And its not clear that you can get out of doing that level of parsing in the kernel anyway (unless you go for a binary namespace representation, and that's a real can of worms). > Is there a difference between these two?: > > long sys_setxattr(char *path, char *name, void *value, size_t size, int flags) > long sys_lsetxattr(char *path, char *name, void *value, size_t size, int flags) > Yes, definately. The easiest reason - there are filesystems which support extended attributes on symlinks already (XFS does), coming from other operating systems, and there should be a way to get at that information too. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Wed Dec 5 22:47:53 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB66lra17507 for linux-xfs-outgoing; Wed, 5 Dec 2001 22:47:53 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB66loo17485 for ; Wed, 5 Dec 2001 22:47:50 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via SMTP id VAA01733 for ; Wed, 5 Dec 2001 21:47:48 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id QAA14844; Thu, 6 Dec 2001 16:46:27 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id QAA50249; Thu, 6 Dec 2001 16:46:26 +1100 (AEDT) Date: Thu, 6 Dec 2001 16:46:25 +1100 From: Nathan Scott To: Anton Altaparmakov Cc: Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [PATCH] Revised extended attributes interface Message-ID: <20011206164625.G50483@wobbly.melbourne.sgi.com> References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <5.1.0.14.2.20011205090142.04ab5b20@pop.cus.cam.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <5.1.0.14.2.20011205090142.04ab5b20@pop.cus.cam.ac.uk>; from aia21@cam.ac.uk on Wed, Dec 05, 2001 at 09:08:12AM +0000 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 09:08:12AM +0000, Anton Altaparmakov wrote: > Looks good to me. Just one tiny point: you seem to like setting error=xyz; > a lot which is completely unnecessary some times. Any particular reason? No compelling reason - I've switched to your version, new patch is here: http://oss.sgi.com/cgi-bin/cvsweb.cgi/linux-2.4-xfs/cmd/xfsmisc/xattr.patch cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Wed Dec 5 23:03:07 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6737Q17904 for linux-xfs-outgoing; Wed, 5 Dec 2001 23:03:07 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6733o17882 for ; Wed, 5 Dec 2001 23:03:03 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via SMTP id WAA10007 for ; Wed, 5 Dec 2001 22:02:50 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id RAA14926; Thu, 6 Dec 2001 17:01:45 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id RAA51622; Thu, 6 Dec 2001 17:01:44 +1100 (AEDT) Date: Thu, 6 Dec 2001 17:01:43 +1100 From: Nathan Scott To: Andrew Ho Cc: Dean Roehrich , linux-xfs@oss.sgi.com Subject: Re: xfsdump error Message-ID: <20011206170143.H50483@wobbly.melbourne.sgi.com> References: <3C0EF677.6B46487F@acm.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <3C0EF677.6B46487F@acm.org>; from andrewho@acm.org on Wed, Dec 05, 2001 at 11:39:19PM -0500 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 11:39:19PM -0500, Andrew Ho wrote: > > Please find the following error. > ... > attr.c:127: `ATTR_OP_IRIX_LIST' undeclared (first use in this function) > You have done a "make install install-dev" in cmd/attr2. Don't do that. Read cmd/attr2/BIG.FAT.WARNING. Do "make install install-dev" in cmd/attr, then go back to building xfsdump and all should be well. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Wed Dec 5 23:39:19 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB67dJB18572 for linux-xfs-outgoing; Wed, 5 Dec 2001 23:39:19 -0800 Received: from smtpzilla5.xs4all.nl (smtpzilla5.xs4all.nl [194.109.127.141]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB67dFo18550 for ; Wed, 5 Dec 2001 23:39:15 -0800 Received: from auto-nb1.xs4all.nl (qn-212-58-167-191.quicknet.nl [212.58.167.191]) by smtpzilla5.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB66crRE017054; Thu, 6 Dec 2001 07:38:53 +0100 (CET) Message-Id: <4.3.2.7.2.20011206073426.035a3d88@pop.xs4all.nl> X-Sender: knuffie@pop.xs4all.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Thu, 06 Dec 2001 07:35:45 +0100 To: Dan Hollis , Simon Matter From: Seth Mos Subject: Re: How to fsck read-only-mounted root filesystem? Cc: linux xfs ml In-Reply-To: References: <3C0E1B6C.F4518BDD@ch.sauter-bc.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 13:40 5-12-2001 -0800, Dan Hollis wrote: >On Wed, 5 Dec 2001, Simon Matter wrote: > > Dan Hollis schrieb: > > > How can one xfs_check/xfs_repair a root filesystem? > > Boot from a bootable CD and perform the repair on the unmounted > > filesystem. > > Have a look at this http://lbt.linuxcare.com/index.epl > >Seems a bit ridiculous to have to install a CDrom just to fsck the root >filesystem. Install what? You don't install anything. You just boot from the CD, fix stuff and then reboot. It takes a bit more time yes, but 5 minutes at most of the corruption is not serious. -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Wed Dec 5 23:49:02 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB67n2c18790 for linux-xfs-outgoing; Wed, 5 Dec 2001 23:49:02 -0800 Received: from anime.net (root@anime.net [63.172.78.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB67mxo18768 for ; Wed, 5 Dec 2001 23:48:59 -0800 Received: from localhost (goemon@localhost) by anime.net (8.11.6/8.9.3) with ESMTP id fB66mqD16649; Wed, 5 Dec 2001 22:48:52 -0800 Date: Wed, 5 Dec 2001 22:48:52 -0800 (PST) From: Dan Hollis To: Seth Mos cc: Simon Matter , linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? In-Reply-To: <4.3.2.7.2.20011206073426.035a3d88@pop.xs4all.nl> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 6 Dec 2001, Seth Mos wrote: > At 13:40 5-12-2001 -0800, Dan Hollis wrote: > >On Wed, 5 Dec 2001, Simon Matter wrote: > > > Dan Hollis schrieb: > > > > How can one xfs_check/xfs_repair a root filesystem? > > > Boot from a bootable CD and perform the repair on the unmounted > > > filesystem. > > > Have a look at this http://lbt.linuxcare.com/index.epl > >Seems a bit ridiculous to have to install a CDrom just to fsck the root > >filesystem. > Install what? You don't install anything. You just boot from the CD, fix > stuff and then reboot. Takes a while to install a cdrom drive... > It takes a bit more time yes, but 5 minutes at most of the corruption is > not serious. Seems a bit silly to have to do it at all, I can see this being a 'show stopper' bug for some. -Dan -- [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Wed Dec 5 23:55:18 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB67tIJ19042 for linux-xfs-outgoing; Wed, 5 Dec 2001 23:55:18 -0800 Received: from relay.xlink.net (relay.xlink.net [193.141.40.5]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB67tDo19020 for ; Wed, 5 Dec 2001 23:55:13 -0800 Received: from lizard.webland.de (lizard.webland.de [194.122.76.201]) by relay.xlink.net (8.9.3/8.8.7) with ESMTP id HAA24238; Thu, 6 Dec 2001 07:55:08 +0100 (MET) Received: (from uucp@localhost) by lizard.webland.de (8.8.8/8.8.7) id HAA19917; Thu, 6 Dec 2001 07:55:08 +0100 (MET) >Received: from mobile.sauter-bc.com (unknown [10.1.6.21]) by basel1.sauter-bc.com (Postfix) with ESMTP id 9A2B857306; Thu, 6 Dec 2001 07:54:44 +0100 (CET) Received: from ch.sauter-bc.com (support.cad.sba [10.1.200.117]) by mobile.sauter-bc.com (Postfix) with ESMTP id 672DA25835; Thu, 6 Dec 2001 07:54:40 +0100 (CET) Message-ID: <3C0F1630.5873C0B6@ch.sauter-bc.com> Date: Thu, 06 Dec 2001 07:54:40 +0100 From: Simon Matter Organization: Sauter AG, Basel X-Mailer: Mozilla 4.77 [de] (X11; U; Linux 2.2.19-6.2.12 i686) X-Accept-Language: de-CH, en MIME-Version: 1.0 To: Dan Hollis Cc: Steve Lord , linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? References: Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Dan Hollis schrieb: > > On 5 Dec 2001, Steve Lord wrote: > > It is also not 'install a cdrom', it is boot with a bootable cdrom. > > This machine doesnt have any cdrom drive, so it is "install a cdrom drive" > just to fix the root fs. :-( But you have a floppy drive, do you? So why don't you take boot/root floppies instead? They are available as well. > > > > After all, reiserfs can reiserfsck a filesystem mounted read-only... > > > Maybe it's a bug in xfs_check/xfs_repair? > > It is deliberate. > > So is changing this behaviour in the "TODO" list, or is it to permanently > remain so? If it is in the TODO list, it will have low priority I guess. There are more important things on the list. > > It was *really* suprising to me that you can't repair a read-only mounted > filesystem, I suspect it will be equally suprising to others. What is life without surprise? Windwos users are also surprised when they learn that we don't have (need) a C: on our computers :-) > > -Dan > -- > [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Thu Dec 6 00:02:59 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB682xu19343 for linux-xfs-outgoing; Thu, 6 Dec 2001 00:02:59 -0800 Received: from anime.net (root@anime.net [63.172.78.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB682to19319 for ; Thu, 6 Dec 2001 00:02:55 -0800 Received: from localhost (goemon@localhost) by anime.net (8.11.6/8.9.3) with ESMTP id fB672bM16901; Wed, 5 Dec 2001 23:02:37 -0800 Date: Wed, 5 Dec 2001 23:02:37 -0800 (PST) From: Dan Hollis To: Simon Matter cc: Steve Lord , linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? In-Reply-To: <3C0F1630.5873C0B6@ch.sauter-bc.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 6 Dec 2001, Simon Matter wrote: > Dan Hollis schrieb: > > On 5 Dec 2001, Steve Lord wrote: > > > It is also not 'install a cdrom', it is boot with a bootable cdrom. > > This machine doesnt have any cdrom drive, so it is "install a cdrom drive" > > just to fix the root fs. :-( > But you have a floppy drive, do you? So why don't you take boot/root > floppies instead? They are available as well. No floppy either. I will have to install one. This sucks :-( > > It was *really* suprising to me that you can't repair a read-only mounted > > filesystem, I suspect it will be equally suprising to others. > What is life without surprise? Windwos users are also surprised when > they learn that we don't have (need) a C: on our computers :-) Suprised in the respect that ext2 and reiserfs have no such limitation, I was expecting xfs to be at least as functional and mature. -Dan -- [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Thu Dec 6 00:21:33 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB68LXq19811 for linux-xfs-outgoing; Thu, 6 Dec 2001 00:21:33 -0800 Received: from mail-nb00s0.nbnet.nb.ca (smtp1.nbnet.nb.ca [198.164.200.23]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB68LRo19789 for ; Thu, 6 Dec 2001 00:21:27 -0800 Received: from wizard ([156.34.217.173]) by mail-nb00s0.nbnet.nb.ca (Post.Office MTA v3.5.3 release 223 ID# 0-72041U145000L145000S0V35) with SMTP id ca; Thu, 6 Dec 2001 03:21:24 -0400 Reply-To: From: "David Wasylciw" To: "Nathan Scott" Cc: Subject: RE: Problems setting ACL with setfacl Date: Thu, 6 Dec 2001 03:21:46 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) In-Reply-To: <20011206112820.B50483@wobbly.melbourne.sgi.com> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Importance: Normal Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Nathan, thanks for the help that got the problem exactly. Any chance of this being tossed in the docs somewhere before the new libacl comes up as I could see others running into it quite easily? Thanks again, Dave -----Original Message----- From: owner-linux-xfs@oss.sgi.com [mailto:owner-linux-xfs@oss.sgi.com]On Behalf Of Nathan Scott Sent: Wednesday, December 05, 2001 8:28 PM To: David Wasylciw Cc: linux-xfs@oss.sgi.com Subject: Re: Problems setting ACL with setfacl hi, On Wed, Dec 05, 2001 at 01:50:51AM -0400, David Wasylciw wrote: > Hello All, > I've recently installed kernel 2.4.16 with the latest XFS all together patch > and of course compiled in ACL support. I'm running Redhat 6.2 and have > installed all the latest acl programs and libraries but for some reason when > I attempt to add a user to the ACL list for a file I get the following > error: > setfacl: file: Resulting ACL > `,user::rw-,user:dwasyl:rw-,mask::---,other::r--': Invalid entry type at > entry 1 This error message is pretty unhelpful - it will get better in the future (when we have a different libacl). > This was generated with the command: setfacl -m u:dwasyl:rw testfile > I checked through the archives and saw something about perhaps needing to > have the permissions listed as rw- but that did not seem to help and > generated the exact same error. > > It would be much appreciated if anyone could help point me in the right > direction and if any more information is needed just let me know. You probably don't have a mask entry in the ACL. Try something like this: $ setfacl -m m::rwx testfile $ setfacl -m u:dwasyl:rw- testfile The above recipe works for me. Andreas' man pages are quite extensive and have a section describing the mask entry - see: http://acl.bestbits.at/man/acl.5.html cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Thu Dec 6 00:47:01 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB68l1Q20305 for linux-xfs-outgoing; Thu, 6 Dec 2001 00:47:01 -0800 Received: from fep02-app.kolumbus.fi (fep02-0.kolumbus.fi [193.229.0.44]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB68kto20283 for ; Thu, 6 Dec 2001 00:46:55 -0800 Received: from there ([62.248.190.29]) by fep02-app.kolumbus.fi (InterMail vM.5.01.03.08 201-253-122-118-108-20010628) with SMTP id <20011206074653.YEND4415.fep02-app.kolumbus.fi@there>; Thu, 6 Dec 2001 09:46:53 +0200 Content-Type: text/plain; charset="iso-8859-1" From: Hristo Grigorov To: Stefan Smietanowski Subject: Re: compilers Date: Thu, 6 Dec 2001 09:48:41 +0200 X-Mailer: KMail [version 1.3.1] Cc: Jean Francois Martinez , Austin Gonyou , Igor Lautar , "'Christian, Chip'" , "Linux XFS (E-mail)" References: <20011205224032.VRAN4415.fep02-app.kolumbus.fi@there> <3C0EB34C.4010904@stesmi.com> In-Reply-To: <3C0EB34C.4010904@stesmi.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <20011206074653.YEND4415.fep02-app.kolumbus.fi@there> Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thursday 06 December 2001 01:52, Stefan Smietanowski wrote: > Hi. > > >>>This is not true. RH 7.1 gcc is not broke(as in can't compile code that > >>>runs), but there was an update to it which fixed certain issues. You > >>>definitely CAN compile kernels with that default gcc though that will > >>>boot. Will you have problems? ymmv. The best course of action is to > >>>either get gcc3, or update your current gcc from updates.redhat.com. > >> > >>Gcc3 is known to NOT work with kernel. > > > > Linux version 2.4.16-xfs (root@magdanoz) (gcc version 3.0.2 20011002 (Red > > Hat Linux 7.1 3.0.1-4)) > > > > Really ? How does it happen that it works here for months already ? Not > > even one single fault. And the system is not idle for sure.... > > It works for you, have you tried installing every single driver in the > kernel tree? You 100% sure you trust all drivers? I'm not talking about > broken code, I'm talking compiler problems. Also, how can you use 3.0.2 > for months already? Was it the ISDN subsystem that didn't work with > gcc3? Something didn't work as it should. Didn't even compile but the > compiler itself died with an internal compiler fault. And lately there > has been kernel modifications to make use of gcc. XFS for one has had a > fix. (Tell me if I'm wrong here). > > I sometimes use gcc3 to compile the kernel also, but that doesn't mean I > trust it on a mission-critical server. > > // Stefan You are may be right... But here I have used kernel with ISDN, XFS and SB Live! support for some months already w/o any problems. It includes all major kernel subsystems also. By using gcc3 I beleive I help open source community to progress ahead. If something goes wrong I will let gcc guys know it immediatelly. Of course that's not good aproach when running "mission-critical" server but I somehow also beleive that XFS will partly save my ass when needed. :))) -- Cheers, Hristo. From owner-linux-xfs@oss.sgi.com Thu Dec 6 00:50:15 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB68oFA20477 for linux-xfs-outgoing; Thu, 6 Dec 2001 00:50:15 -0800 Received: from smtpzilla1.xs4all.nl (smtpzilla1.xs4all.nl [194.109.127.137]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB68oBo20452 for ; Thu, 6 Dec 2001 00:50:11 -0800 Received: from auto-nb1.xs4all.nl (coltex.xs4all.nl [213.84.127.168]) by smtpzilla1.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB67o0sk045763; Thu, 6 Dec 2001 08:50:00 +0100 (CET) Message-Id: <4.3.2.7.2.20011206084345.02e5c618@pop.xs4all.nl> X-Sender: knuffie@pop.xs4all.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Thu, 06 Dec 2001 08:46:53 +0100 To: Dan Hollis , Xianglong Yuan From: Seth Mos Subject: Re: Follow up -- Re: Files on XFS not safe?! Cc: In-Reply-To: References: <20011205140358.F3446@real-uma.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 14:16 5-12-2001 -0800, Dan Hollis wrote: >On Wed, 5 Dec 2001, Xianglong Yuan wrote: > > I just cross-by an excellent article on various journaling > > techniques used in ext3. > > http://www-106.ibm.com/developerworks/linux/library/l-fs8.html > >It looks like tux2 is taking a different approach, and one of the goals is >total data integrity -- not just metadata. They are using phase trees. >It's been described as "failsafe" whereas XFS and most other journaling fs >are just "crash resistant". I don't know if your database would like that approach. XFS makes sure my database stays intact after a crash while the database itself (Progress 9) manages the rest of the recovery process for stuff that it was working in while it crashed. Tux2 seems to be a really good idea for static content systems but I am afraid that databases will not like phase trees. I have not read the page yet so I might be all wrong ofcourse. Cheers -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Thu Dec 6 01:13:50 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB69Don20970 for linux-xfs-outgoing; Thu, 6 Dec 2001 01:13:50 -0800 Received: from smtpzilla2.xs4all.nl (smtpzilla2.xs4all.nl [194.109.127.138]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB69Djo20948 for ; Thu, 6 Dec 2001 01:13:45 -0800 Received: from auto-nb1.xs4all.nl (coltex.xs4all.nl [213.84.127.168]) by smtpzilla2.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB68DZjx043206; Thu, 6 Dec 2001 09:13:35 +0100 (CET) Message-Id: <4.3.2.7.2.20011206090743.02e72ed0@pop.xs4all.nl> X-Sender: knuffie@pop.xs4all.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Thu, 06 Dec 2001 09:10:28 +0100 To: Dan Hollis From: Seth Mos Subject: Re: How to fsck read-only-mounted root filesystem? Cc: Simon Matter , linux xfs ml In-Reply-To: References: <4.3.2.7.2.20011206073426.035a3d88@pop.xs4all.nl> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 22:48 5-12-2001 -0800, Dan Hollis wrote: >On Thu, 6 Dec 2001, Seth Mos wrote: > > At 13:40 5-12-2001 -0800, Dan Hollis wrote: > > >On Wed, 5 Dec 2001, Simon Matter wrote: > > > > Dan Hollis schrieb: > > > > > How can one xfs_check/xfs_repair a root filesystem? > > > > Boot from a bootable CD and perform the repair on the unmounted > > > > filesystem. > > > > Have a look at this http://lbt.linuxcare.com/index.epl > > >Seems a bit ridiculous to have to install a CDrom just to fsck the root > > >filesystem. > > Install what? You don't install anything. You just boot from the CD, fix > > stuff and then reboot. > >Takes a while to install a cdrom drive... I always carry one around when I am going to maintain a server. I have a server at a colo which doesn't have a cdrom drive either. I just take one with me when I go there. I am forced to take the machine out of the rack anyways if I am going do something to it. But fixing a corrupt fs while the system is online is not reassuring for me either. > > It takes a bit more time yes, but 5 minutes at most of the corruption is > > not serious. > >Seems a bit silly to have to do it at all, I can see this being a 'show >stopper' bug for some. For some yes. -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Thu Dec 6 01:30:42 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB69UgX21339 for linux-xfs-outgoing; Thu, 6 Dec 2001 01:30:42 -0800 Received: from minnie.omroep.nl (minnie.omroep.nl [145.58.30.4]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB69UZo21315 for ; Thu, 6 Dec 2001 01:30:36 -0800 Received: (from smap@localhost) by minnie.omroep.nl (SGI-8.9.3/8.9.3) id JAA52818 for ; Thu, 6 Dec 2001 09:30:28 +0100 (CET) Received: from nos-smtp.nos.nl (145.58.12.125 "HELO nos-smtp.nos.nl") by minnie.omroep.nl with SMTP (smap v3.0 nederlandse publieke omroep) id xma4961862; Thu, 6 Dec 01 09:30:27 +0100 Received: FROM exchange.nos.nl BY nos-smtp.nos.nl ; Thu Dec 06 09:41:34 2001 +0100 Received: by EXCHANGE with Internet Mail Service (5.5.2650.21) id ; Thu, 6 Dec 2001 09:31:50 +0100 Message-ID: <35F87783B600D411A1430060943F469A589731@EXCHANGE> From: Matthijs van der Klip To: "'ivanr@sgi.com'" , Matthijs van der Klip Cc: Linux XFS Mailing List Subject: RE: using xfsdump to synchronise filesystems? Date: Thu, 6 Dec 2001 09:31:41 +0100 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Disposition: inline Content-Type: text/plain Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 6 Dec 2001, Ivan Rayner wrote: > I don't think xfsdump is a very good choice here. xfsdump will scan the > entire filesystem every time you run it, even if you select a higher level > dump. I doubt that for a 10GB filesystem it will be able to run to > completion within 5 minutes ('course this depends on your circumstances). Ah, this is information I needed. I already wondered if xfsdump did a scan or something more intelligent. BTW I can do a level 2 xfsdump of a 20Gb filesystem in around 80 secs. This was timed on a filesystem with around 300.000 files. > I think you'd be better off monitoring the ftp log to see what files have > changed and then use rsync on those. I'm sure it'd be fairly easy to do > this in something like perl. Problem with this is the information contained in the ftp logs itself by default is not sufficient: 1) I can use the xferlog, but this contains only xfers, no mkdir's, chmod's etc. Moreover spaces in filenames are replaced by underscores so the log essentialy contains invalid filenames. 2) I can use the ftp command log, but this contains literal ftp commands, which I would have to parse in order to filter the information I need. It doesn't even contains full pathnames to files... Ofcourse this could be solved by hacking the ftp daemon so it writes the log I want it to, but I'm not convinced this is the way to go. Best regards, Matthijs van der Klip NOS Dutch Public Broadcasting Organisation [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Thu Dec 6 01:43:47 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB69hlf21638 for linux-xfs-outgoing; Thu, 6 Dec 2001 01:43:47 -0800 Received: from mxzilla4.xs4all.nl (mxzilla4.xs4all.nl [194.109.6.48]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB69heo21612 for ; Thu, 6 Dec 2001 01:43:41 -0800 Received: from xs3.xs4all.nl (dirk@xs3.xs4all.nl [194.109.6.44]) by mxzilla4.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB68hcdI071230 for ; Thu, 6 Dec 2001 09:43:38 +0100 (CET) Received: (from dirk@localhost) by xs3.xs4all.nl (8.9.0/8.9.0) id JAA16446 for linux-xfs@oss.sgi.com; Thu, 6 Dec 2001 09:43:38 +0100 (CET) Date: Thu, 6 Dec 2001 09:43:37 +0100 From: Dirk Jan To: xfs ml Subject: dma troubles Message-ID: <20011206094337.B12776@xs4all.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi all. I am having some troubles with DMA timeouts. This is a preliminary report, I am at work and don't have messages and such. I found some, probably related, messages when searching for "dma timeout", one of the messages pointed to a thread on linux-kernel (and seemed related to a specific chipset) The system is a P133 compaq with 2 HD's. Distribution is debian woody, kernel is 2.4.14. The 2nd HD (hdc) contains the installion, the 1st HD (hda) is the new HD to which I want to copy the filesystem. Moving files on 1 HD is no problem but from hdc to hda gives dma timeouts after a while. (somthing like cp -a / /mnt/root where root is a mounted partition from hda) When is goes wrong the ide drive isn't available anymore. It seems as if the partition table is gone but it looks like the kernel ignores /dev/hda. Anyway, I blaimed xfs for this but when I made the filesystem ext2 I still get these dma timeouts. Then I read in archive about broken or soon to be broken HD's I got a little worried. The 1fst HD is a brand new one. So, final test: I tarred the whole installation on hdc ftp'ed to another pc, ftp'ed it backed to hda and untarred. No errors at all. So I guess it has something to do with accessing two ide channels at the same time, but I haven't something like this before. Oke, if anyone has a clue about what is going on please explain/clearify to this me. If error messages are needed (and it would explain things better I guess) I won't have time to make them before tomorrow night) If you made it so far, thanks for reading. -- Dirk From owner-linux-xfs@oss.sgi.com Thu Dec 6 01:55:54 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB69tsq22306 for linux-xfs-outgoing; Thu, 6 Dec 2001 01:55:54 -0800 Received: from zomeronline.net (root@cc14521-a.groni1.gr.nl.home.com [217.120.199.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB69tmo22283 for ; Thu, 6 Dec 2001 01:55:48 -0800 Received: from aldertlaptop (imenz07.biol.rug.nl [129.125.132.247]) by zomeronline.net (8.11.6/8.11.6) with SMTP id fB68tZp02409; Thu, 6 Dec 2001 09:55:35 +0100 Message-ID: <004001c17e33$92fbec60$f7847d81@aldertlaptop> From: "Aldert Zomer" To: "Matthijs van der Klip" , Cc: "Linux XFS Mailing List" References: <35F87783B600D411A1430060943F469A589731@EXCHANGE> Subject: Re: using xfsdump to synchronise filesystems? Date: Thu, 6 Dec 2001 09:54:10 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4807.1700 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk ----- Original Message ----- From: "Matthijs van der Klip" To: ; "Matthijs van der Klip" Cc: "Linux XFS Mailing List" Sent: 06 December, 2001 09:31 Subject: RE: using xfsdump to synchronise filesystems? > > I think you'd be better off monitoring the ftp log to see what files have > > changed and then use rsync on those. I'm sure it'd be fairly easy to do > > this in something like perl. Perhaps Intermezzo is an alternative? Snippet from their HOWTO: ---------- InterMezzo uses an existing disk file system as the storage location for all data. At present we support ext3, but soon also ReiserFS and XFS might be supported. When an ext3 formatted disk volume is mounted with file system type InterMezzo instead of ext3, the InterMezzo software starts managing all access to the file system. It keeps the logs of modification records and negotiates permits to modify the disk file system, to avoid conflicting updates during connected operation. ----------- At present they do not support XFS, however they might do in the future. Intermezzo can be found here: http://www.inter-mezzo.org/ . Aldert Zomer -- Ing. Aldert Zomer, | Tel: +31 50 36 32 360 Research Technician, | Fax: +31 50 36 35 205 IMEnz Bioengineering, | Email: zomeral@biol.rug.nl PO BOX 14, 9750 AA Haren, Netherlands | Homepage: http://www.imenz.com ------------------------------------------------------------------------ Automated Medline search: http://molgen.biol.rug.nl/cgi-bin/biomail/users.pl Personal Webpage: http://www.zomeronline.net/ From owner-linux-xfs@oss.sgi.com Thu Dec 6 02:48:01 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6Am1M28412 for linux-xfs-outgoing; Thu, 6 Dec 2001 02:48:01 -0800 Received: from mail.loewe-komp.de (mail.loewe-komp.de [62.156.155.230]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6Aluo28390 for ; Thu, 6 Dec 2001 02:47:57 -0800 Received: from loewe-komp.de (pippin [192.168.169.19]) by mail.loewe-komp.de (8.11.0/8.11.0/SuSE Linux 8.11.0-0.4) with ESMTP id fB69noS23452; Thu, 6 Dec 2001 10:49:50 +0100 X-Authentication-Warning: mail.loewe-komp.de: Host pippin [192.168.169.19] claimed to be loewe-komp.de Message-ID: <3C0F3EC8.506A243D@loewe-komp.de> Date: Thu, 06 Dec 2001 10:47:52 +0100 From: Peter =?iso-8859-1?Q?W=E4chtler?= Organization: LOEWE. Hannover X-Mailer: Mozilla 4.78 [de] (X11; U; Linux 2.4.9-ac3 i686) X-Accept-Language: de, en MIME-Version: 1.0 To: Dan Hollis CC: Harri Haataja , Xianglong Yuan , Paulo Sergio Lemes Queiroz , linux-xfs@oss.sgi.com Subject: Re: Files on XFS not safe?! References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Dan Hollis schrieb: > > On Wed, 5 Dec 2001, Harri Haataja wrote: > > On Wed, Dec 05, 2001 at 10:47:41AM -0500, Xianglong Yuan wrote: > > > >-Paulo Sergio Lemes Queiroz [05 Dec 2001 13:15 -0200] wrote: > > > > http://oss.sgi.com/projects/xfs/faq.html#nulls > > > Thanks. Is this true for all the journaling FS, say ReiserFS, > > > JFS? > > Maybe I might bring up this > > http://kerneltrap.com/article.php?sid=389 > > FWIW I never had reiserfs nulling files... it is either more resistant to > such situations or it operates in a totally different way than XFS. > But perhaps "scrambled tails"? AFAIK, ReiserFS moves the tails of files to put several into one disk block. I lost a lot of work due to this - ok it was just setiathome :)) From owner-linux-xfs@oss.sgi.com Thu Dec 6 02:56:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6Au3p28637 for linux-xfs-outgoing; Thu, 6 Dec 2001 02:56:03 -0800 Received: from anime.net (root@anime.net [63.172.78.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6Au0o28615 for ; Thu, 6 Dec 2001 02:56:00 -0800 Received: from localhost (goemon@localhost) by anime.net (8.11.6/8.9.3) with ESMTP id fB69tar19372; Thu, 6 Dec 2001 01:55:36 -0800 Date: Thu, 6 Dec 2001 01:55:36 -0800 (PST) From: Dan Hollis To: Peter =?iso-8859-1?Q?W=E4chtler?= cc: Harri Haataja , Xianglong Yuan , Paulo Sergio Lemes Queiroz , Subject: Re: Files on XFS not safe?! In-Reply-To: <3C0F3EC8.506A243D@loewe-komp.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=ISO-8859-1 X-MIME-Autoconverted: from 8bit to quoted-printable by anime.net id fB69tar19372 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id fB6Au0o28616 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 6 Dec 2001, Peter Wächtler wrote: > Dan Hollis schrieb: > > FWIW I never had reiserfs nulling files... it is either more resistant to > > such situations or it operates in a totally different way than XFS. > But perhaps "scrambled tails"? AFAIK, ReiserFS moves the tails of files to put > several into one disk block. Never had scrambled tails either.. > I lost a lot of work due to this - ok it was just setiathome :)) You can turn off tailmerging with mount option 'notail' ... -Dan -- [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Thu Dec 6 02:58:54 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6AwsM28808 for linux-xfs-outgoing; Thu, 6 Dec 2001 02:58:54 -0800 Received: from gusi.leathercollection.ph (gusi.leathercollection.ph [202.163.192.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6Awmo28786 for ; Thu, 6 Dec 2001 02:58:49 -0800 Received: from localhost (localhost [127.0.0.1]) by gusi.leathercollection.ph (Postfix) with ESMTP id BBB9EC00B61 for ; Thu, 6 Dec 2001 17:58:45 +0800 (PHT) Received: from mail.leathercollection.ph (gusi.leathercollection.ph [192.168.0.1]) by gusi.leathercollection.ph (Postfix) with ESMTP id 9F84DC00B60 for ; Thu, 6 Dec 2001 17:58:43 +0800 (PHT) Date: Thu, 6 Dec 2001 17:58:43 +0800 (PHT) From: Federico Sevilla III To: Linux XFS Mailing List Subject: Re: Files on XFS not safe?! In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by AMaViS perl-11 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, 5 Dec 2001 at 13:49, Dan Hollis wrote: > FWIW I never had reiserfs nulling files... it is either more resistant > to such situations or it operates in a totally different way than XFS. Both ext2 and ReiserFS I find are less prone, but not immune. And I've had messups with ReiserFS, too. They don't get nulled out, though. I'd just get some random character in my newly-edited config file (ie: a "W" at a line by itself somewhere). If you want your data secure in boxes that nosedive regularly you can either use ext3's full data journalling or one of the BSDs which do not mount their filesystems asynchronously. I find, though, that for servers that are locked up, pre-tested, and backed up by a UPS, XFS is the best filesystem available on Linux performance-wise and stability-wise. :) --> Jijo -- Federico Sevilla III :: jijo@leathercollection.ph Network Administrator :: The Leather Collection, Inc. GnuPG Key: From owner-linux-xfs@oss.sgi.com Thu Dec 6 03:00:36 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6B0a229009 for linux-xfs-outgoing; Thu, 6 Dec 2001 03:00:36 -0800 Received: from k-7.stesmi.com (IDENT:root@as4-1-7.has.s.bonet.se [217.215.31.238]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6B0Ro28987 for ; Thu, 6 Dec 2001 03:00:28 -0800 Received: from stesmi.com (voyager.stesmi.com [192.168.1.11]) by k-7.stesmi.com (8.11.2/8.8.7) with ESMTP id fB69xbG02194; Thu, 6 Dec 2001 10:59:37 +0100 Message-ID: <3C0F4202.9010908@stesmi.com> Date: Thu, 06 Dec 2001 11:01:38 +0100 From: Stefan Smietanowski User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.6) Gecko/20011120 X-Accept-Language: en-us MIME-Version: 1.0 To: Hristo Grigorov CC: Jean Francois Martinez , Austin Gonyou , Igor Lautar , "'Christian, Chip'" , "Linux XFS (E-mail)" Subject: Re: compilers References: <20011205224032.VRAN4415.fep02-app.kolumbus.fi@there> <3C0EB34C.4010904@stesmi.com> <20011206074653.YEND4415.fep02-app.kolumbus.fi@there> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hristo Grigorov wrote: > On Thursday 06 December 2001 01:52, Stefan Smietanowski wrote: > >>Hi. >> >> >>>>>This is not true. RH 7.1 gcc is not broke(as in can't compile code that >>>>>runs), but there was an update to it which fixed certain issues. You >>>>>definitely CAN compile kernels with that default gcc though that will >>>>>boot. Will you have problems? ymmv. The best course of action is to >>>>>either get gcc3, or update your current gcc from updates.redhat.com. >>>>> >>>>Gcc3 is known to NOT work with kernel. >>>> >>>Linux version 2.4.16-xfs (root@magdanoz) (gcc version 3.0.2 20011002 (Red >>>Hat Linux 7.1 3.0.1-4)) >>> >>>Really ? How does it happen that it works here for months already ? Not >>>even one single fault. And the system is not idle for sure.... >>> >>It works for you, have you tried installing every single driver in the >>kernel tree? You 100% sure you trust all drivers? I'm not talking about >>broken code, I'm talking compiler problems. Also, how can you use 3.0.2 >>for months already? Was it the ISDN subsystem that didn't work with >>gcc3? Something didn't work as it should. Didn't even compile but the >>compiler itself died with an internal compiler fault. And lately there >>has been kernel modifications to make use of gcc. XFS for one has had a >>fix. (Tell me if I'm wrong here). >> >>I sometimes use gcc3 to compile the kernel also, but that doesn't mean I >>trust it on a mission-critical server. >> >>// Stefan >> > > You are may be right... But here I have used kernel with ISDN, XFS and SB Live! > support for some months already w/o any problems. It includes all major kernel > subsystems also. > > By using gcc3 I beleive I help open source community to progress ahead. If something > goes wrong I will let gcc guys know it immediatelly. Of course that's not good aproach > when running "mission-critical" server but I somehow also beleive that XFS will partly > save my ass when needed. :))) > > Sure, I know, that's the reason I use gcc3 also but it was the "stable" part I was commenting on, nothing else. // Stefan From owner-linux-xfs@oss.sgi.com Thu Dec 6 04:18:39 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6CIdo03433 for linux-xfs-outgoing; Thu, 6 Dec 2001 04:18:39 -0800 Received: from s1.uklinux.net (mail.uklinux.net [80.84.72.21]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6CIXo03411 for ; Thu, 6 Dec 2001 04:18:33 -0800 Received: from pyewacket.nic.uklinux.net (host213-122-7-126.btconnect.com [213.122.7.126]) (authenticated) by s1.uklinux.net (8.11.6/8.11.6) with ESMTP id fB6BIOo11125 for ; Thu, 6 Dec 2001 11:18:24 GMT Envelope-To: Received: from [127.0.0.1] (helo=there) by pyewacket.nic.uklinux.net with smtp (Exim 3.33 #1) id 16BwX1-0001iy-00 for linux-xfs@oss.sgi.com; Thu, 06 Dec 2001 11:17:35 +0000 Content-Type: text/plain; charset="iso-8859-1" From: nic To: linux xfs ml Subject: [Offtopic] Re: How to fsck read-only-mounted root filesystem? Date: Thu, 6 Dec 2001 11:17:35 +0000 X-Mailer: KMail [version 1.3.1] References: <4.3.2.7.2.20011206073426.035a3d88@pop.xs4all.nl> <4.3.2.7.2.20011206090743.02e72ed0@pop.xs4all.nl> In-Reply-To: <4.3.2.7.2.20011206090743.02e72ed0@pop.xs4all.nl> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk > At 22:48 5-12-2001 -0800, Dan Hollis wrote: > >Suprised in the respect that ext2 and reiserfs have no such limitation, I > >was expecting xfs to be at least as functional and mature. > >Seems a bit silly to have to do it at all, I can see this being a 'show > >stopper' bug for some. You may find this a bit odd, but commercial Unices[1] have often had quirky means to repair the essential filesystems and more dependencies at boot time. But then they came on better firmware (eg. OpenBoot) so you could set up alternative devices, etc. Steve mentioned that Irix (and Solaris, and I think AIX) all use miniroot, under which one mounts the old filesystems and fixes them or upgrades the operating system, etc. The LBT (and indeed if you don't need the full power/convenience/vendor-neutraility of lbt, the SGI installation CD) give you the same means of fixing your root fs as big, expensive Unix boxes have used for years. It may not be as convenient for you, but it works and has been used on some of the most expensive Unix machines in the world - your implication that is XFS immature given its roots and these facts would seem to be an inaccurate remark. Perhaps you should consider setting up a spare disk (or possibly just a partition) from which you can boot with a root partition on it. This would save you installing a CD drive. Ordinary x86 BIOSes don't provide much help for when the primary HD fails (though obviously some do) so better hardware will make this easier. [Any technical inaccuracies not related to Linux or XFS in this mail should be sent to me off-list]. nic Footnotes: [1] OK, Irix and Solaris - I've only had minimal experience with a few of the others. From owner-linux-xfs@oss.sgi.com Thu Dec 6 06:20:55 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6EKtK05533 for linux-xfs-outgoing; Thu, 6 Dec 2001 06:20:55 -0800 Received: from porsta.cs.Helsinki.FI (root@porsta.cs.Helsinki.FI [128.214.48.124]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6EKoo05500 for ; Thu, 6 Dec 2001 06:20:50 -0800 Received: from melkki.cs.Helsinki.FI (sslwrap@localhost [127.0.0.1]) by porsta.cs.Helsinki.FI (8.11.6/8.11.6) with ESMTP id fB6DKiE17057 for ; Thu, 6 Dec 2001 15:20:46 +0200 Received: (from hhaataja@localhost) by melkki.cs.Helsinki.FI (8.11.6/8.11.2) id fB6DKcI06430 for linux-xfs@oss.sgi.com; Thu, 6 Dec 2001 15:20:38 +0200 Date: Thu, 6 Dec 2001 15:20:38 +0200 From: Harri Haataja Cc: linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? Message-ID: <20011206152038.A6273@cs.helsinki.fi> References: <3C0F1630.5873C0B6@ch.sauter-bc.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from goemon@anime.net on Wed, Dec 05, 2001 at 11:02:37PM -0800 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 11:02:37PM -0800, Dan Hollis wrote: > On Thu, 6 Dec 2001, Simon Matter wrote: > > Dan Hollis schrieb: > > > On 5 Dec 2001, Steve Lord wrote: > > > > It is also not 'install a cdrom', it is boot with a bootable cdrom. > > > This machine doesnt have any cdrom drive, so it is "install a cdrom drive" > > > just to fix the root fs. :-( > > But you have a floppy drive, do you? So why don't you take boot/root > > floppies instead? They are available as well. > > No floppy either. I will have to install one. This sucks :-( I would call only one method of booting the computer "all eggs in one frail basket" (and even that only if I was feeling very polite). Other boot methods would be a second hard drive you can plug in, network boot and another miniroot (which won't work in case the bootblock (or MBR) gets hosed). > > > It was *really* suprising to me that you can't repair a read-only mounted > > > filesystem, I suspect it will be equally suprising to others. > > What is life without surprise? Windwos users are also surprised when > > they learn that we don't have (need) a C: on our computers :-) > > Suprised in the respect that ext2 and reiserfs have no such limitation, I > was expecting xfs to be at least as functional and mature. Different is not neccessarily.. oh, hell. I'm not going there =) -- When you need a helpline for breakfast cereals, it's time to start thinking about tearing down civilisation and giving the ants a go. -- Chris King From owner-linux-xfs@oss.sgi.com Thu Dec 6 06:20:24 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6EKOm05463 for linux-xfs-outgoing; Thu, 6 Dec 2001 06:20:24 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6EKIo05439 for ; Thu, 6 Dec 2001 06:20:18 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via SMTP id FAA08826 for ; Thu, 6 Dec 2001 05:20:15 -0800 (PST) mail_from (ivanr@sgi.com) Received: from omen.melbourne.sgi.com (omen.melbourne.sgi.com [134.14.55.139]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id AAA16866; Fri, 7 Dec 2001 00:18:59 +1100 From: ivanr@sgi.com Received: from localhost (ivanr@localhost) by omen.melbourne.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id AAA66799; Fri, 7 Dec 2001 00:18:58 +1100 (EST) X-Authentication-Warning: omen.melbourne.sgi.com: ivanr owned process doing -bs Date: Fri, 7 Dec 2001 00:18:58 +1100 X-X-Sender: ivanr@omen.melbourne.sgi.com To: Matthijs van der Klip cc: Linux XFS Mailing List Subject: RE: using xfsdump to synchronise filesystems? In-Reply-To: <35F87783B600D411A1430060943F469A589731@EXCHANGE> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 6 Dec 2001, Matthijs van der Klip wrote: > On Thu, 6 Dec 2001, Ivan Rayner wrote: > > I think you'd be better off monitoring the ftp log to see what files have > > changed and then use rsync on those. I'm sure it'd be fairly easy to do > > this in something like perl. > > Problem with this is the information contained in the ftp logs itself by > default is not sufficient: > > 1) I can use the xferlog, but this contains only xfers, no mkdir's, chmod's > etc. Moreover spaces in filenames are replaced by underscores so the log > essentialy contains invalid filenames. > > 2) I can use the ftp command log, but this contains literal ftp commands, > which I would have to parse in order to filter the information I need. It > doesn't even contains full pathnames to files... > > Ofcourse this could be solved by hacking the ftp daemon so it writes the log > I want it to, but I'm not convinced this is the way to go. Another idea... IRIX has thing called fam (file aleteration monitor) which can notify apps when files or directories are changed. fam has been ported to Linux, so you could possibly write a small program that uses the fam API and can start rsyncs as files are changed or added, etc. See http://oss.sgi.com/projects/fam/ fam consists of a useland daemon called fam, and a kernel patch called imon (inode monitor). I'm not sure what the current state of the kernel patch is, but it might be worth seeing if you can get this going. fam/imon are not filesystem specific. There's a link from the fam faq to a project that looks just like what you want: http://www.nas.nasa.gov/Groups/WWW/subpages/topology.html HTH, Ivan -- Ivan Rayner ivanr@sgi.com From owner-linux-xfs@oss.sgi.com Thu Dec 6 06:25:13 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6EPDC05824 for linux-xfs-outgoing; Thu, 6 Dec 2001 06:25:13 -0800 Received: from blr.vsnl.net.in (blr.vsnl.net.in [202.54.12.6]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6EP9o05802 for ; Thu, 6 Dec 2001 06:25:10 -0800 Received: from mail (PPP-200-7-219.bng.vsnl.net.in [203.200.7.219]) by blr.vsnl.net.in (Postfix) with SMTP id 88DCD7B1E for ; Thu, 6 Dec 2001 18:53:46 +0530 (IST) Received: from 192.9.200.146 [192.9.200.146] by mail (1.61/SMTPD) at Sat, 17 Nov 01 20:16:07 PST Message-ID: <3C0EBE9E.9050904@uiscpl.com> Date: Wed, 05 Dec 2001 19:41:02 -0500 From: iqbal Reply-To: iqbal@uiscpl.com Organization: U & I system design limited User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.4) Gecko/20010914 X-Accept-Language: en-us MIME-Version: 1.0 To: linux-xfs@oss.sgi.com Subject: kernel debugger Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Gateway: --->> 1 - POP3/SMTP Gateway 5.0 for GIFTmail <<--- Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, I have read in some article saying that, sgi xfs for linux has kernel debugger, So can i some information regarding the use of kernel debugger in xfs, thanking you iqbal From owner-linux-xfs@oss.sgi.com Thu Dec 6 06:33:32 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6EXWj06075 for linux-xfs-outgoing; Thu, 6 Dec 2001 06:33:32 -0800 Received: from main.braxis.co.uk (root@main.braxis.co.uk [213.77.40.29]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6EXOo06051 for ; Thu, 6 Dec 2001 06:33:25 -0800 Received: (from kszysiu@localhost) by main.braxis.co.uk (8.11.6/8.11.6) id fB6DWsh20719; Thu, 6 Dec 2001 14:32:54 +0100 Date: Thu, 6 Dec 2001 14:32:54 +0100 From: Krzysztof Rusocki To: iqbal Cc: linux-xfs@oss.sgi.com Subject: Re: kernel debugger Message-ID: <20011206143253.A20373@main.braxis.co.uk> References: <3C0EBE9E.9050904@uiscpl.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <3C0EBE9E.9050904@uiscpl.com>; from iqbal@uiscpl.com on Wed, Dec 05, 2001 at 07:41:02PM -0500 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 07:41:02PM -0500, iqbal wrote: > > Hi, > > I have read in some article saying that, sgi xfs for linux has kernel > debugger, > So can i some information regarding the use of kernel debugger in xfs, Hello iqbal, Kernel debugger (kdb) is an SGI OpenSource Project, you can find info about it on http://oss.sgi.com/projects/kdb/ but as they seem a little bit out of date (i think) - read linux/Documentation/kdb/* from linux-xfs source tree. KDB patches for mainstream kernels are avaialable on ftp://oss.sgi.com/projects/kdb/ (afair). Cheers, Krzysztof > > thanking you > iqbal > From owner-linux-xfs@oss.sgi.com Thu Dec 6 06:56:36 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6EuaV06754 for linux-xfs-outgoing; Thu, 6 Dec 2001 06:56:36 -0800 Received: from turtle.lila.net (pD4B9F486.dip.t-dialin.net [212.185.244.134]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6EuSo06732 for ; Thu, 6 Dec 2001 06:56:29 -0800 Received: from mpg.goe.ni.schule.de (localhost [127.0.0.1]) by turtle.lila.net (8.11.6/8.11.6/SuSE Linux 0.5) with ESMTP id fB6DiB101985 for ; Thu, 6 Dec 2001 14:44:16 +0100 Message-ID: <3C0F762B.A6E4C25D@mpg.goe.ni.schule.de> Date: Thu, 06 Dec 2001 14:44:11 +0100 From: "Jan H. Schrewe" Reply-To: jschrewe@mpg.goe.ni.schule.de X-Mailer: Mozilla 4.78 [de] (X11; U; Linux 2.4.12-xfs i686) X-Accept-Language: en MIME-Version: 1.0 To: linux-xfs@oss.sgi.com Subject: [Fwd: using xfsdump to synchronise filesystems?] Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk (Should also go to the list .....) "Jan H. Schrewe" schrieb: > > Matthijs van der Klip schrieb: > > > > On Thu, 6 Dec 2001, Ivan Rayner wrote: > > > I don't think xfsdump is a very good choice here. xfsdump will scan the > > > entire filesystem every time you run it, even if you select a higher level > > > dump. I doubt that for a 10GB filesystem it will be able to run to > > > completion within 5 minutes ('course this depends on your circumstances). > > > > Ah, this is information I needed. I already wondered if xfsdump did a scan > > or something more intelligent. > > > > BTW I can do a level 2 xfsdump of a 20Gb filesystem in around 80 secs. This > > was timed on a filesystem with around 300.000 files. > > > > > I think you'd be better off monitoring the ftp log to see what files have > > > changed and then use rsync on those. I'm sure it'd be fairly easy to do > > > this in something like perl. > > > > > > > > Problem with this is the information contained in the ftp logs itself by > > default is not sufficient: > > > > 1) I can use the xferlog, but this contains only xfers, no mkdir's, chmod's > > etc. Moreover spaces in filenames are replaced by underscores so the log > > essentialy contains invalid filenames. > > > > 2) I can use the ftp command log, but this contains literal ftp commands, > > which I would have to parse in order to filter the information I need. It > > doesn't even contains full pathnames to files... > > > > Ofcourse this could be solved by hacking the ftp daemon so it writes the log > > I want it to, but I'm not convinced this is the way to go. > > > > > > > > Best regards, > > Is it correct that only one server recives the data and the others should just > mirror the files? Then you should have a look at DRBD (it does raid-1 over > ethernet) at > http://www.complang.tuwien.ac.at/reisner/drbd/ > > It was developed for High Availibilty clusters but I think it should do waht you > want. > > cheers > > Jan > > > Matthijs van der Klip > > NOS Dutch Public Broadcasting Organisation > > > > [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Thu Dec 6 07:12:57 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6FCvR07271 for linux-xfs-outgoing; Thu, 6 Dec 2001 07:12:57 -0800 Received: from ahriman.bucharest.roedu.net (ahriman.bucharest.roedu.net [141.85.128.71]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6FCqo07249 for ; Thu, 6 Dec 2001 07:12:52 -0800 Received: (qmail 25763 invoked by uid 1000); 6 Dec 2001 14:15:40 -0000 Received: from localhost (sendmail-bs@127.0.0.1) by localhost with SMTP; 6 Dec 2001 14:15:40 -0000 Date: Thu, 6 Dec 2001 16:15:40 +0200 (EET) From: Mihai RUSU X-X-Sender: To: nic cc: linux xfs ml Subject: Re: [Offtopic] Re: How to fsck read-only-mounted root filesystem? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi i he could create a separate parition with a "mini system" on it and configure/run lilo to boot from it On Thu, 6 Dec 2001, nic wrote: > Ordinary x86 BIOSes don't provide much help for when the primary HD fails > (though obviously some do) so better hardware will make this easier. > > [Any technical inaccuracies not related to Linux or XFS in this mail should > be sent to me off-list]. > well, newer BIOSes permit to access the SETUP tool on the serial console, pretty cool huh ? ---------------------------- Mihai RUSU "... and what if this is as good as it gets ?" From owner-linux-xfs@oss.sgi.com Thu Dec 6 07:22:34 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6FMY507554 for linux-xfs-outgoing; Thu, 6 Dec 2001 07:22:34 -0800 Received: from k-7.stesmi.com (IDENT:root@as4-1-7.has.s.bonet.se [217.215.31.238]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6FMSo07532 for ; Thu, 6 Dec 2001 07:22:28 -0800 Received: from stesmi.com (voyager.stesmi.com [192.168.1.11]) by k-7.stesmi.com (8.11.2/8.8.7) with ESMTP id fB6ELSG04191; Thu, 6 Dec 2001 15:21:28 +0100 Message-ID: <3C0F7F65.7040809@stesmi.com> Date: Thu, 06 Dec 2001 15:23:33 +0100 From: Stefan Smietanowski User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.6) Gecko/20011120 X-Accept-Language: en-us MIME-Version: 1.0 To: Dan Hollis CC: Simon Matter , Steve Lord , linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? References: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Dan Hollis wrote: > On Thu, 6 Dec 2001, Simon Matter wrote: > >>Dan Hollis schrieb: >> >>>On 5 Dec 2001, Steve Lord wrote: >>> >>>>It is also not 'install a cdrom', it is boot with a bootable cdrom. >>>> >>>This machine doesnt have any cdrom drive, so it is "install a cdrom drive" >>>just to fix the root fs. :-( >>> >>But you have a floppy drive, do you? So why don't you take boot/root >>floppies instead? They are available as well. >> > > No floppy either. I will have to install one. This sucks :-( > > >>>It was *really* suprising to me that you can't repair a read-only mounted >>>filesystem, I suspect it will be equally suprising to others. >>> >>What is life without surprise? Windwos users are also surprised when >>they learn that we don't have (need) a C: on our computers :-) >> > > Suprised in the respect that ext2 and reiserfs have no such limitation, I > was expecting xfs to be at least as functional and mature. It's just designed differently. If You go buy a ferrari and go whine that you can't fit that piano in the trunk, then you have to think what's important for you. Fitting the piano or driving fast. If repair-on-mounted-filesystem is what you're after then you use a filesystem that supports it. If you're after a good, fast and stable (and mature) filesystem you stick to xfs. The choice is yours. // Stefan From owner-linux-xfs@oss.sgi.com Thu Dec 6 07:46:32 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6FkWV07956 for linux-xfs-outgoing; Thu, 6 Dec 2001 07:46:32 -0800 Received: from mail.dkp.com ([204.191.16.3]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6FkSo07927 for ; Thu, 6 Dec 2001 07:46:28 -0800 Received: from ranma.dkp.com (ranma.dkp.com [205.150.40.12]) by mail.dkp.com (Postfix) with ESMTP id 2EBC91AB32 for ; Thu, 6 Dec 2001 09:46:27 -0500 (EST) Received: by ranma.dkp.com (Postfix, from userid 168) id B65065AA0B; Thu, 6 Dec 2001 09:46:26 -0500 (EST) Date: Thu, 6 Dec 2001 09:46:26 -0500 From: Andrew Klaassen To: linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? Message-ID: <20011206094626.B5971@dkp.com> Mail-Followup-To: linux xfs ml References: <1007591316.30179.10.camel@jen.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1007591316.30179.10.camel@jen.americas.sgi.com> User-Agent: Mutt/1.3.23i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Wed, Dec 05, 2001 at 04:28:36PM -0600, Steve Lord wrote: > > > Dan Hollis schrieb: > > > > How can one xfs_check/xfs_repair a root filesystem? > No, it is just that the environment xfs comes from (Irix) > tends to use other mechanisms to fix root filesystems - like > an alternate root, or a miniroot. > > It is also not 'install a cdrom', it is boot with a bootable > cdrom. > > Doing a user space fixup of a live filesystem also relies upon > the block device cache and the filesystem's cache of metadata > being coherent. If they are not then this approach does not > work and can lead to nasty crashes very quickly. XFS was until > recently using completely different caches for metadata than > the block interface, they now use the same memory - mostly. > Until they are coherent there is absolutely no point in doing > modifications to a mounted XFS fs from user space. Huh. I could've sworn that I've xfs_repaired a filesystem mounted readonly in the past... Andrew Klaassen - must be experiencing brain bit-rot... From owner-linux-xfs@oss.sgi.com Thu Dec 6 07:56:52 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6FuqC08214 for linux-xfs-outgoing; Thu, 6 Dec 2001 07:56:52 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6Fumo08192 for ; Thu, 6 Dec 2001 07:56:48 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id GAA03616 for ; Thu, 6 Dec 2001 06:56:50 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id IAA3768759; Thu, 6 Dec 2001 08:55:30 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id IAA66126; Thu, 6 Dec 2001 08:55:29 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB6Et0t07858; Thu, 6 Dec 2001 08:55:00 -0600 Subject: RE: using xfsdump to synchronise filesystems? From: Steve Lord To: Matthijs van der Klip Cc: "'ivanr@sgi.com'" , Linux XFS Mailing List In-Reply-To: <35F87783B600D411A1430060943F469A589731@EXCHANGE> References: <35F87783B600D411A1430060943F469A589731@EXCHANGE> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 06 Dec 2001 08:54:59 -0600 Message-Id: <1007650499.5749.6.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 2001-12-06 at 02:31, Matthijs van der Klip wrote: > On Thu, 6 Dec 2001, Ivan Rayner wrote: > > I don't think xfsdump is a very good choice here. xfsdump will scan the > > entire filesystem every time you run it, even if you select a higher level > > dump. I doubt that for a 10GB filesystem it will be able to run to > > completion within 5 minutes ('course this depends on your circumstances). > > Ah, this is information I needed. I already wondered if xfsdump did a scan > or something more intelligent. > > BTW I can do a level 2 xfsdump of a 20Gb filesystem in around 80 secs. This > was timed on a filesystem with around 300.000 files. XFS does have a special interface to walk the inodes in the filesystem without walking through all the directories, this is called bulkstat and it is used by dump. This is why it can scan the filesystem so fast. Steve -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Thu Dec 6 08:24:27 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6GORC08866 for linux-xfs-outgoing; Thu, 6 Dec 2001 08:24:27 -0800 Received: from starship.berlin (dsl-213-023-038-110.arcor-ip.net [213.23.38.110]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6GOKo08843 for ; Thu, 6 Dec 2001 08:24:20 -0800 Received: from daniel by starship.berlin with local (Exim 3.32 #1 (Debian)) id 16C0PD-0000ot-00; Thu, 06 Dec 2001 16:25:47 +0100 Content-Type: text/plain; charset="iso-8859-1" From: Daniel Phillips To: Nathan Scott , Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher Subject: Re: [PATCH] Revised extended attributes interface Date: Thu, 6 Dec 2001 16:25:41 +0100 X-Mailer: KMail [version 1.3.2] Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <20011206164131.F50483@wobbly.melbourne.sgi.com> In-Reply-To: <20011206164131.F50483@wobbly.melbourne.sgi.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On December 6, 2001 06:41 am, Nathan Scott wrote: > hey there. > > > I still don't like the class parsing inside the kernel, it's hard to see > > what is good about that. > > I guess it ultimately comes down to simplicity. The IRIX interfaces > have this separation of name and namespace - each operation has to > indicate which namespace is to be used. That becomes very messy when > you wish to work with multiple attribute names and namespaces at once. > Since the namespace is intimately tied to the name anyway, this idea > of specifying the two components together provides very clean APIs. Right now we have two namespaces, user and system. That's one bit of information, and the proposal is to represent it with 5-7 bytes, passing it on every call, and decoding it with a memcmp or similar. This is just extra fluff as far as I can see, and provides every bit as much opportunity for implementing a private API as the original cmd parameter did, by encoding whatever one pleases before the dot. > The term "parsing" is a bit of an overstatement too. We're talking > strncmp() complexity here, not lex/yacc. ;) And its not clear that > you can get out of doing that level of parsing in the kernel anyway > (unless you go for a binary namespace representation, and that's a > real can of worms). I'm suggesting we take a look at that. > > Is there a difference between these two?: > > > > long sys_setxattr(char *path, char *name, void *value, size_t size, int flags) > > long sys_lsetxattr(char *path, char *name, void *value, size_t size, int flags) > > > > Yes, definately. The easiest reason - there are filesystems which > support extended attributes on symlinks already (XFS does), coming > from other operating systems, and there should be a way to get at > that information too. OK, well it looks like you're going a little overboard here in dividing out the functionality. What you're talking about is 'follow symlink or not', right? That really does sound to me as though it's naturally expressed with a flag bit. I really don't see a compelling reason to go beyond 8 syscalls: get, fget, set, fset, del, fdel, list, flist -- Daniel From owner-linux-xfs@oss.sgi.com Thu Dec 6 10:15:21 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6IFLe22997 for linux-xfs-outgoing; Thu, 6 Dec 2001 10:15:21 -0800 Received: from fnal.gov (heffalump.fnal.gov [131.225.9.20]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6IFHo22971 for ; Thu, 6 Dec 2001 10:15:17 -0800 Received: from vk-clued0.fnal.gov ([131.225.224.23]) by smtp.fnal.gov (PMDF V6.0-24 #37519) with ESMTP id <0GNX003A4MLFQW@smtp.fnal.gov> for linux-xfs@oss.sgi.com; Thu, 06 Dec 2001 11:15:15 -0600 (CST) Date: Thu, 06 Dec 2001 11:15:15 -0600 (CST) From: vkuznet Subject: XFS and quota problem To: linux-xfs@oss.sgi.com Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, I have a problem with quota command on XFS partitions. We have a cluster of ~150 nodes with XFS fs and after upgrading to 2.4.14 kernel our quota -v refuse to work. The source RPM for quota package from SGI (found on rpmfind.net) compiled successfully but show 0 as your quota limit. Example: > vkuznet@vk-clued0(12:40:37)> usr/bin/quota -v > Disk quotas for user vkuznet (uid 1212): > Filesystem blocks quota limit grace files quota limit > grace > /d02ka.fnal.gov:/desktop/home5/vkuznet > 0 0 0 0 0 0 > /plum-clued0:/disk2/fnal/ups > 0 0 0 0 0 0 > /plum-clued0.fnal.gov:/disk2/local > 0 0 0 0 0 0 We do have a variety of disks (local, cross mounted from Linux and from SGI file server) and none of them show valid values for quotas. Can anyone explain the problem and/or provide or point me to valid patch. I'm not on XFS maling list (sorry too many mails) and will appreciate if you replay to my Email directly. Thanks, Valentin. From owner-linux-xfs@oss.sgi.com Thu Dec 6 10:12:53 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6ICrg22890 for linux-xfs-outgoing; Thu, 6 Dec 2001 10:12:53 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6ICeo22866 for ; Thu, 6 Dec 2001 10:12:41 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id SAA1580732 for ; Thu, 6 Dec 2001 18:12:33 +0100 (CET) mail_from (sandeen@sgi.com) Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3543860; Thu, 6 Dec 2001 11:11:16 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id LAA49877; Thu, 6 Dec 2001 11:11:15 -0600 (CST) Subject: Re: patch anaconda XFS error From: Eric Sandeen To: =?ISO-8859-1?Q?J=E9r=F4me?= Tournier Cc: linux-xfs@oss.sgi.com In-Reply-To: <20011204161225.A8486@etoile.magasin1.auchan.com> References: <20011204161225.A8486@etoile.magasin1.auchan.com> Content-Type: text/plain; charset=ISO-8859-1 X-Mailer: Evolution/0.99.2 (Preview Release) Date: 06 Dec 2001 11:04:37 -0600 Message-Id: <1007658278.13482.5.camel@stout.americas.sgi.com> Mime-Version: 1.0 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id fB6ICjo22868 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi Jérôme - Sorry for the late reply! Can you tell me the exact steps you took that ended in this error? I have not been able to reproduce it. Thanks, -Eric On Tue, 2001-12-04 at 09:12, Jérôme Tournier wrote: > Hello, > i use the anaconda package with XFS patch. > I have already wrote here because when generating a new distribution, the installation proccess stopped because it can't load the xfs modules (depedencies problem). > I still not resolved this problem (altought i add xfs_support and pagebuf in mk-images.i386 script). > When i don't have this problem, i can continue but what we know found strange is that when installing the system in graphical mode, > everything ok, but when installing with a boot disk, i have this error message: > Traceback (innermost last): > File "/usr/bin/anaconda", line 620, in ? > intf.run(id, dispatch, configFileData) > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/text.py", line 386, in run > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 1004, in __call__ > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 948, in editCb > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 657, in editPartitionRequest > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 486, in fsOptionsDialog > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 251, in makeFsList > File "/usr/lib/python1.5/snack.py", line 100, in setCurrent > KeyError: > > Local variables in innermost frame: > self: > item: > > > Any idea ? > Thanks > -- > Jérôme -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Thu Dec 6 10:50:16 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6IoG026316 for linux-xfs-outgoing; Thu, 6 Dec 2001 10:50:16 -0800 Received: from etoile.infra.idealx.com (sete.idealx.com [213.41.87.82]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6Io3o26287 for ; Thu, 6 Dec 2001 10:50:03 -0800 Received: (qmail 14307 invoked by uid 500); 6 Dec 2001 17:49:43 -0000 From: "Jérôme Tournier" Date: Thu, 6 Dec 2001 18:49:43 +0100 To: Eric Sandeen Cc: linux-xfs@oss.sgi.com Subject: Re: patch anaconda XFS error Message-ID: <20011206184943.A2023@etoile.magasin.auchan.com> References: <20011204161225.A8486@etoile.magasin1.auchan.com> <1007658278.13482.5.camel@stout.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <1007658278.13482.5.camel@stout.americas.sgi.com> User-Agent: Mutt/1.3.23i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Le Thu, Dec 06, 2001 at 11:04:37AM -0600, Eric Sandeen à écrit: # Hi Jérôme - # # Sorry for the late reply! # # Can you tell me the exact steps you took that ended in this error? I # have not been able to reproduce it. Here is the histrorical problem. I used first the SGI CDROM modified distribution. I install the system in text mode and i have the error message i give the mail before. Then i do another installation in graphical mode: everything was fine ! I then tried to rebuild the distribution using the anaconda-7.2-11XFS.src.rpm package. To generate the distribution, i use the following command: cd /usr/lib/anaconda-runtime ./genhdlist /home/ftp/pub/redhat/ ./buildinstall /home/ftp/pub/redhat/ > /tmp/buildinstall.log 2>&1 One time over 3 or 5, when installing the distribution, i have error messages because of dependencies problem so that the xfs module can't be loaded. Here is the error: xfs_support: pagdepmod: *** Unresolved symbols in fdomain_cs But as i said, sometimes it's ok!! After, i added the reiserfs support to anaconda. In the file fsset.py, i defined this like: def fileSystemTypeGetDefault(): if fileSystemTypeGet('xfs').isSupported(): return fileSystemTypeGet('xfs') elif fileSystemTypeGet('ext3').isSupported(): return fileSystemTypeGet('ext3') elif fileSystemTypeGet('reiserfs').isSupported(): return fileSystemTypeGet('reiserfs') elif fileSystemTypeGet('ext2').isSupported(): return fileSystemTypeGet('ext2') else: raise ValueError, "You have neither xfs, ext3, nor ext2 support in your kernel!" and i changed the section reative to reiserfs class reiserfsFileSystem(FileSystemType): def __init__(self): FileSystemType.__init__(self) self.partedFileSystemType = parted.file_system_type_get("reiserfs") self.formattable = 1 self.checked = 1 self.linuxnativefs = 1 self.supported = 1 self.name = "reiserfs" self.maxSize = 2 * 1024 * 1024 I then have again the same error messages as above. When i invert the first line of fileSystemTypeGetDefault to have def fileSystemTypeGetDefault(): if fileSystemTypeGet('ext3').isSupported(): return fileSystemTypeGet('ext3') elif fileSystemTypeGet('xfs').isSupported(): return fileSystemTypeGet('xfs') elif fileSystemTypeGet('reiserfs').isSupported(): return fileSystemTypeGet('reiserfs') elif fileSystemTypeGet('ext2').isSupported(): return fileSystemTypeGet('ext2') else: raise ValueError, "You have neither xfs, ext3, nor ext2 support in your kernel!" i don't have anymore the error message, but i still have the dependencie problem. Strange, isn't it ? Am i enought clear ? Do you want a patch of the fsset.py modified file ? I have another question: i want to incorporate the LVM support to anaconda. Did you have already work on this ? Do you know if there's something available (patch) ? Thanks -- Jérôme # On Tue, 2001-12-04 at 09:12, Jérôme Tournier wrote: # > Hello, # > i use the anaconda package with XFS patch. # > I have already wrote here because when generating a new distribution, the installation proccess stopped because it can't load the xfs modules (depedencies problem). # > I still not resolved this problem (altought i add xfs_support and pagebuf in mk-images.i386 script). # > When i don't have this problem, i can continue but what we know found strange is that when installing the system in graphical mode, # > everything ok, but when installing with a boot disk, i have this error message: # > Traceback (innermost last): # > File "/usr/bin/anaconda", line 620, in ? # > intf.run(id, dispatch, configFileData) # > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/text.py", line 386, in run # > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 1004, in __call__ # > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 948, in editCb # > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 657, in editPartitionRequest # > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 486, in fsOptionsDialog # > File "/var/tmp/anaconda-7.2//usr/lib/anaconda/textw/partition_text.py", line 251, in makeFsList # > File "/usr/lib/python1.5/snack.py", line 100, in setCurrent # > KeyError: # > # > Local variables in innermost frame: # > self: # > item: # > # > # > Any idea ? # > Thanks # > -- # > Jérôme # -- # Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs # sandeen@sgi.com SGI, Inc. -- Jérôme From owner-linux-xfs@oss.sgi.com Thu Dec 6 10:56:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6Iu3p26634 for linux-xfs-outgoing; Thu, 6 Dec 2001 10:56:03 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6Itso26612 for ; Thu, 6 Dec 2001 10:55:55 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id SAA1621103 for ; Thu, 6 Dec 2001 18:55:49 +0100 (CET) mail_from (sandeen@sgi.com) Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id LAA3785196; Thu, 6 Dec 2001 11:54:33 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id LAA85728; Thu, 6 Dec 2001 11:54:32 -0600 (CST) Subject: Re: patch anaconda XFS error From: Eric Sandeen To: =?ISO-8859-1?Q?J=E9r=F4me?= Tournier Cc: linux-xfs@oss.sgi.com In-Reply-To: <20011206184943.A2023@etoile.magasin.auchan.com> References: <20011204161225.A8486@etoile.magasin1.auchan.com> <1007658278.13482.5.camel@stout.americas.sgi.com> <20011206184943.A2023@etoile.magasin.auchan.com> Content-Type: text/plain; charset=ISO-8859-1 X-Mailer: Evolution/0.99.2 (Preview Release) Date: 06 Dec 2001 11:47:54 -0600 Message-Id: <1007660874.13468.7.camel@stout.americas.sgi.com> Mime-Version: 1.0 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id fB6Itvo26613 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 2001-12-06 at 11:49, Jérôme Tournier wrote: > Le Thu, Dec 06, 2001 at 11:04:37AM -0600, Eric Sandeen à écrit: > # Hi Jérôme - > # > # Sorry for the late reply! > # > # Can you tell me the exact steps you took that ended in this error? I > # have not been able to reproduce it. > > Here is the histrorical problem. I used first the SGI CDROM modified distribution. > I install the system in text mode and i have the error message i give the mail before. And when did this error occur? After what steps in the install? re: adding reiserfs and support, I have not tried either, and it's not really on my todo list, so I'm afraid I may not be able to offer much help. -Eric -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Thu Dec 6 11:07:16 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6J7G027035 for linux-xfs-outgoing; Thu, 6 Dec 2001 11:07:16 -0800 Received: from etoile.infra.idealx.com (sete.idealx.com [213.41.87.82]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6J7Bo27013 for ; Thu, 6 Dec 2001 11:07:11 -0800 Received: (qmail 15412 invoked by uid 500); 6 Dec 2001 18:06:50 -0000 From: "Jérôme Tournier" Date: Thu, 6 Dec 2001 19:06:50 +0100 To: Eric Sandeen Cc: =?iso-8859-1?B?Suly9G1l?= Tournier , linux-xfs@oss.sgi.com Subject: Re: patch anaconda XFS error Message-ID: <20011206190650.B2023@etoile.magasin.auchan.com> References: <20011204161225.A8486@etoile.magasin1.auchan.com> <1007658278.13482.5.camel@stout.americas.sgi.com> <20011206184943.A2023@etoile.magasin.auchan.com> <1007660874.13468.7.camel@stout.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <1007660874.13468.7.camel@stout.americas.sgi.com> User-Agent: Mutt/1.3.23i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Le Thu, Dec 06, 2001 at 11:47:54AM -0600, Eric Sandeen à écrit: # On Thu, 2001-12-06 at 11:49, Jérôme Tournier wrote: # > Le Thu, Dec 06, 2001 at 11:04:37AM -0600, Eric Sandeen à écrit: # > # Hi Jérôme - # > # # > # Sorry for the late reply! # > # # > # Can you tell me the exact steps you took that ended in this error? I # > # have not been able to reproduce it. # > # > Here is the histrorical problem. I used first the SGI CDROM modified distribution. # > I install the system in text mode and i have the error message i give the mail before. # # And when did this error occur? After what steps in the install? Just when selecting the filesystem i want to use -- Jérôme From owner-linux-xfs@oss.sgi.com Thu Dec 6 11:08:12 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6J8Cg27173 for linux-xfs-outgoing; Thu, 6 Dec 2001 11:08:12 -0800 Received: from mailhost.idcomm.com (mailhost.idcomm.com [207.40.196.14]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6J84o27150 for ; Thu, 6 Dec 2001 11:08:04 -0800 Received: from idcomm.com (IDENT:nV7Jei3jgbu9QVbAllPVdcTBx0YPdDJP@k56-pip57.idcomm.com [209.60.72.184]) by mailhost.idcomm.com (8.10.2/8.10.0) with ESMTP id fB6I9aE21588 for ; Thu, 6 Dec 2001 11:09:36 -0700 Message-ID: <3C0FB448.9D6ECA02@idcomm.com> Date: Thu, 06 Dec 2001 11:09:12 -0700 From: "D. Stimits" Reply-To: stimits@idcomm.com X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.6-pre1-xfs-4 i686) X-Accept-Language: en MIME-Version: 1.0 CC: "Linux XFS (E-mail)" Subject: Re: compilers References: <20011205224032.VRAN4415.fep02-app.kolumbus.fi@there> <3C0EB34C.4010904@stesmi.com> <20011206074653.YEND4415.fep02-app.kolumbus.fi@there> <3C0F4202.9010908@stesmi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Stefan Smietanowski wrote: > > Hristo Grigorov wrote: > > > On Thursday 06 December 2001 01:52, Stefan Smietanowski wrote: > > > >>Hi. > >> > >> > >>>>>This is not true. RH 7.1 gcc is not broke(as in can't compile code that > >>>>>runs), but there was an update to it which fixed certain issues. You > >>>>>definitely CAN compile kernels with that default gcc though that will > >>>>>boot. Will you have problems? ymmv. The best course of action is to > >>>>>either get gcc3, or update your current gcc from updates.redhat.com. > >>>>> > >>>>Gcc3 is known to NOT work with kernel. > >>>> > >>>Linux version 2.4.16-xfs (root@magdanoz) (gcc version 3.0.2 20011002 (Red > >>>Hat Linux 7.1 3.0.1-4)) > >>> > >>>Really ? How does it happen that it works here for months already ? Not > >>>even one single fault. And the system is not idle for sure.... > >>> > >>It works for you, have you tried installing every single driver in the > >>kernel tree? You 100% sure you trust all drivers? I'm not talking about > >>broken code, I'm talking compiler problems. Also, how can you use 3.0.2 > >>for months already? Was it the ISDN subsystem that didn't work with > >>gcc3? Something didn't work as it should. Didn't even compile but the > >>compiler itself died with an internal compiler fault. And lately there > >>has been kernel modifications to make use of gcc. XFS for one has had a > >>fix. (Tell me if I'm wrong here). > >> > >>I sometimes use gcc3 to compile the kernel also, but that doesn't mean I > >>trust it on a mission-critical server. > >> > >>// Stefan > >> > > > > You are may be right... But here I have used kernel with ISDN, XFS and SB Live! > > support for some months already w/o any problems. It includes all major kernel > > subsystems also. > > > > By using gcc3 I beleive I help open source community to progress ahead. If something > > goes wrong I will let gcc guys know it immediatelly. Of course that's not good aproach > > when running "mission-critical" server but I somehow also beleive that XFS will partly > > save my ass when needed. :))) > > > > > > Sure, I know, that's the reason I use gcc3 also but it was the "stable" > part I was commenting on, nothing else. > > // Stefan One other thing that needs to be tested (perhaps more than the others) before blessing the compiler for all uses would be every possible optimization. It seems that various optimizations are where a lot of the failures occur, e.g., compiling (or coding for features of) for i386, i686, or Athlon. It seems that a number of places where one compiler or another breaks things (or does things correctly but the app expects old broken behavior from) are optimizations that rearrange the asm incorrectly under one level of optimizing, versus others. I have to wonder just how enormous of a project it would be to create a compiler testbed that tries every possible code combination and optimization and then reports on deviations between what is expected or what is in a standard, versus what it got (such a project could require a large cluster of both computers and people). D. Stimits, stimits@idcomm.com From owner-linux-xfs@oss.sgi.com Thu Dec 6 11:14:01 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6JE1727408 for linux-xfs-outgoing; Thu, 6 Dec 2001 11:14:01 -0800 Received: from speedycat.clubphoto.local (mail-gw.clubphoto.com [64.124.34.66]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6JDuo27386 for ; Thu, 6 Dec 2001 11:13:56 -0800 Received: from clubphoto.com (localhost.localdomain [127.0.0.1]) by speedycat.clubphoto.local (8.11.2/8.11.2) with ESMTP id fB6I7Cb16958 for ; Thu, 6 Dec 2001 10:07:12 -0800 Message-ID: <3C0FB3D0.B23E6185@clubphoto.com> Date: Thu, 06 Dec 2001 10:07:12 -0800 From: "Drew 'Cheetah!' Maxwell" X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.2-SGI_XFS_1.0 i586) X-Accept-Language: en MIME-Version: 1.0 To: linux-xfs@oss.sgi.com Subject: Linux w/XFS on Athlon systems Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Fellow XFS users: -------------------------------------------------------------------------------- My company has been working with XFS-Linux on Intel systems for some time and has had some difficulties getting a stable Linux w/XFS release to work on some of our older K6-2 systems. Instead of butting our heads against a case of potentially suspect hardware, we're looking to upgrade to a group of Athlon-based systems. In this, I have a few questions: 1. What can those of you running Linux w/XFS recommend for motherboards to go with an Athlon processor? (We might as well get hardware we know works by experience.) 2. Are they any special changes to your kernels used to work correctly with an Athlon-based system? Our initial attempts at doing an Athlon + Linux + XFS solution haven't worked. We've tried the following methods: 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using gcc 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using kgcc Pre-patched XFS 1.0.2 kernel source from SGI, compiled using kgcc Any assistance is greatly appreciated. --Drew 'Cheetah!' Maxwell [cheetah@clubphoto.com] ClubPhoto (http://www.clubphoto.com) Pictures, Prints, Gifts, and More! From owner-linux-xfs@oss.sgi.com Thu Dec 6 11:24:35 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6JOZh27728 for linux-xfs-outgoing; Thu, 6 Dec 2001 11:24:35 -0800 Received: from c1880033-a.chmpgn1.il.home.com (c1880033-a.chmpgn1.il.home.com [24.36.232.71]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6JOTo27705 for ; Thu, 6 Dec 2001 11:24:29 -0800 Received: from jesse (helo=localhost) by c1880033-a.chmpgn1.il.home.com with local-esmtp (Exim 3.33 #1 (Debian)) id 16C3B1-0007b6-00; Thu, 06 Dec 2001 12:23:19 -0600 Date: Thu, 6 Dec 2001 12:23:19 -0600 (CST) From: Jesse Hall X-X-Sender: jdhall@espresso To: "Drew 'Cheetah!' Maxwell" cc: linux-xfs@oss.sgi.com Subject: Re: Linux w/XFS on Athlon systems In-Reply-To: <3C0FB3D0.B23E6185@clubphoto.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk What do you mean by "haven't worked"? I've been using a Tyan Thunder K7 with dual Athlon MPs with XFS (tracking CVS) for a few months with great reliability and stability. HTH, Jesse On Thu, 6 Dec 2001, Drew 'Cheetah!' Maxwell wrote: > Fellow XFS users: > -------------------------------------------------------------------------------- > My company has been working with XFS-Linux on Intel systems for some > time > and has had some difficulties getting a stable Linux w/XFS release to > work > on some of our older K6-2 systems. Instead of butting our heads against > a > case of potentially suspect hardware, we're looking to upgrade to a > group > of Athlon-based systems. In this, I have a few questions: > > 1. What can those of you running Linux w/XFS recommend for motherboards > to > go with an Athlon processor? > (We might as well get hardware we know works by experience.) > 2. Are they any special changes to your kernels used to work correctly > with > an Athlon-based system? > > Our initial attempts at doing an Athlon + Linux + XFS solution haven't > worked. > We've tried the following methods: > 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using > gcc > 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using > kgcc > Pre-patched XFS 1.0.2 kernel source from SGI, compiled using kgcc > > Any assistance is greatly appreciated. > > --Drew 'Cheetah!' Maxwell [cheetah@clubphoto.com] > ClubPhoto (http://www.clubphoto.com) Pictures, Prints, Gifts, and More! > > From owner-linux-xfs@oss.sgi.com Thu Dec 6 11:33:31 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6JXV127982 for linux-xfs-outgoing; Thu, 6 Dec 2001 11:33:31 -0800 Received: from chimta02 (chimta02.algx.net [216.99.233.77]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6JXOo27959 for ; Thu, 6 Dec 2001 11:33:24 -0800 Received: from jtsdell (66-2-81-28.customer.algx.net [66.2.81.28]) by chimmx02.algx.net (iPlanet Messaging Server 5.1 (built May 7 2001)) with ESMTP id <0GNX00DP3Q7MWZ@chimmx02.algx.net> for linux-xfs@oss.sgi.com; Thu, 06 Dec 2001 12:33:23 -0600 (CST) Date: Thu, 06 Dec 2001 13:31:17 -0500 (EST) From: jtrostel@snapserver.com Subject: Re: Linux w/XFS on Athlon systems In-reply-to: To: Jesse Hall Cc: linux-xfs@oss.sgi.com, "Drew 'Cheetah!' Maxwell" Reply-to: jtrostel@snapserver.com Message-id: Organization: Quantum Corp. / NASD MIME-version: 1.0 X-Mailer: XFMail 1.5.1 on Linux Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7BIT X-Priority: 3 (Normal) Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Multiple that answer by two ;-> You need to be more specific in your description of the failure mode. On 06-Dec-2001 Jesse Hall wrote: > What do you mean by "haven't worked"? > > I've been using a Tyan Thunder K7 with dual Athlon MPs with XFS (tracking > CVS) for a few months with great reliability and stability. > > HTH, > Jesse > > On Thu, 6 Dec 2001, Drew 'Cheetah!' Maxwell wrote: >> Fellow XFS users: >> ----------------------------------------------------------------------------- >> --- >> My company has been working with XFS-Linux on Intel systems for some >> time >> and has had some difficulties getting a stable Linux w/XFS release to >> work >> on some of our older K6-2 systems. Instead of butting our heads against >> a >> case of potentially suspect hardware, we're looking to upgrade to a >> group >> of Athlon-based systems. In this, I have a few questions: >> >> 1. What can those of you running Linux w/XFS recommend for motherboards >> to >> go with an Athlon processor? >> (We might as well get hardware we know works by experience.) >> 2. Are they any special changes to your kernels used to work correctly >> with >> an Athlon-based system? >> >> Our initial attempts at doing an Athlon + Linux + XFS solution haven't >> worked. >> We've tried the following methods: >> 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using >> gcc >> 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using >> kgcc >> Pre-patched XFS 1.0.2 kernel source from SGI, compiled using kgcc >> >> Any assistance is greatly appreciated. >> >> --Drew 'Cheetah!' Maxwell [cheetah@clubphoto.com] >> ClubPhoto (http://www.clubphoto.com) Pictures, Prints, Gifts, and More! >> >> -- John M. Trostel Senior Software Engineer Quantum Corp. / NASD jtrostel@snapserver.com From owner-linux-xfs@oss.sgi.com Thu Dec 6 11:36:37 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6JabU28286 for linux-xfs-outgoing; Thu, 6 Dec 2001 11:36:37 -0800 Received: from speedycat.clubphoto.local (mail-gw.clubphoto.com [64.124.34.66]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6JaYo28263 for ; Thu, 6 Dec 2001 11:36:34 -0800 Received: from clubphoto.com (localhost.localdomain [127.0.0.1]) by speedycat.clubphoto.local (8.11.2/8.11.2) with ESMTP id fB6ITeb16973; Thu, 6 Dec 2001 10:29:40 -0800 Message-ID: <3C0FB914.51EED80B@clubphoto.com> Date: Thu, 06 Dec 2001 10:29:40 -0800 From: "Drew 'Cheetah!' Maxwell" X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.2-SGI_XFS_1.0 i586) X-Accept-Language: en MIME-Version: 1.0 To: Jesse Hall CC: linux-xfs@oss.sgi.com Subject: Re: Linux w/XFS on Athlon systems References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Jesse Hall wrote: -------------------------------------------------------------------------------- > > What do you mean by "haven't worked"? > > I've been using a Tyan Thunder K7 with dual Athlon MPs with XFS (tracking > CVS) for a few months with great reliability and stability. > By "haven't worked", I mean generating a Kernel Oops, and never being able to finish bringing the system up to multiuser mode. (It'll die somewhere in the midst of the interactive menus as it brings up daemons.) --Drew 'Cheetah!' Maxwell [cheetah@clubphoto.com] ClubPhoto (http://www.clubphoto.com) Pictures, Prints, Gifts, and More! From owner-linux-xfs@oss.sgi.com Thu Dec 6 11:43:33 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6JhXC28670 for linux-xfs-outgoing; Thu, 6 Dec 2001 11:43:33 -0800 Received: from eclectic.kluge.net (IDENT:root@dsl092-071-242.bos1.dsl.speakeasy.net [66.92.71.242]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6JhSo28648 for ; Thu, 6 Dec 2001 11:43:29 -0800 Received: (from felicity@localhost) by eclectic.kluge.net (8.11.6/8.11.6) id fB6IhR000358 for linux-xfs@oss.sgi.com; Thu, 6 Dec 2001 13:43:27 -0500 Date: Thu, 6 Dec 2001 13:43:27 -0500 From: Theo Van Dinter To: linux-xfs@oss.sgi.com Subject: Re: Linux w/XFS on Athlon systems Message-ID: <20011206134327.S31315@kluge.net> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from jtrostel@snapserver.com on Thu, Dec 06, 2001 at 01:31:17PM -0500 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, Dec 06, 2001 at 01:31:17PM -0500, jtrostel@snapserver.com wrote: > Multiple that answer by two ;-> > > You need to be more specific in your description of the failure mode. > > >> 1. What can those of you running Linux w/XFS recommend for motherboards > >> to > >> go with an Athlon processor? > >> (We might as well get hardware we know works by experience.) > >> 2. Are they any special changes to your kernels used to work correctly > >> with > >> an Athlon-based system? Add another to the "it works" group. I've got an Athlon 850 on a KA7-100 motherboard. It's been working reliably with XFS for months (and now with LVM too). Standard RH7.2 install with the XFS admin apps and kernel RPM installed. (both the athlon and 686 kernel work fine.) Does non-XFS installs work on your Athlons, or is it specifically when you use the XFS-enabled kernels, or ...? -- Randomly Generated Tagline: "Robotic tape changer mistook operator's tie for a backup tape." - Today's BOFH Excuse From owner-linux-xfs@oss.sgi.com Thu Dec 6 11:56:48 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6Jum529092 for linux-xfs-outgoing; Thu, 6 Dec 2001 11:56:48 -0800 Received: from crunch.mw.mediaone.net ([24.30.48.217]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6Juco29069 for ; Thu, 6 Dec 2001 11:56:38 -0800 Received: from squash.mw.mediaone.net (squash.mw.mediaone.net [192.168.1.16]) by crunch.mw.mediaone.net (Postfix) with ESMTP id CF0735446; Thu, 6 Dec 2001 13:56:35 -0500 (EST) Subject: Re: Linux w/XFS on Athlon systems From: Joe Landman To: "Drew 'Cheetah!' Maxwell" Cc: linux-xfs@oss.sgi.com In-Reply-To: <3C0FB3D0.B23E6185@clubphoto.com> References: <3C0FB3D0.B23E6185@clubphoto.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.13 (Preview Release) Date: 06 Dec 2001 13:49:41 -0500 Message-Id: <1007664581.7654.61.camel@squash.mw.mediaone.net> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Drew: I would go with a distribution that supports XFS. That would be Mandrake 8.1 or MSC.Linux from MSC.Software (I work the them). Both work quite well on Athlon hardware (Tyan, Abit, Asus, Gigabyte, ECS). RedHat 7.2 can be shoehorned to take XFS as can be seen from the excellent work of the folks at SGI who redid the RedHat installer. It is a shame the folks over at RedHat havent seen the light. Do the standard diagnostic stuff on your hardware. AMD Athlon + Motherboard + Linux + XFS does work quite well. Only problems I have had seems to have stemmed from devfs. A devfs=nomount seems to make most of them go away. Joe On Thu, 2001-12-06 at 13:07, Drew 'Cheetah!' Maxwell wrote: > Fellow XFS users: > -------------------------------------------------------------------------------- > My company has been working with XFS-Linux on Intel systems for some > time > and has had some difficulties getting a stable Linux w/XFS release to > work > on some of our older K6-2 systems. Instead of butting our heads against > a > case of potentially suspect hardware, we're looking to upgrade to a > group > of Athlon-based systems. In this, I have a few questions: > > 1. What can those of you running Linux w/XFS recommend for motherboards > to > go with an Athlon processor? > (We might as well get hardware we know works by experience.) > 2. Are they any special changes to your kernels used to work correctly > with > an Athlon-based system? > > Our initial attempts at doing an Athlon + Linux + XFS solution haven't > worked. > We've tried the following methods: > 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using > gcc > 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using > kgcc > Pre-patched XFS 1.0.2 kernel source from SGI, compiled using kgcc > > Any assistance is greatly appreciated. > > --Drew 'Cheetah!' Maxwell [cheetah@clubphoto.com] > ClubPhoto (http://www.clubphoto.com) Pictures, Prints, Gifts, and More! From owner-linux-xfs@oss.sgi.com Thu Dec 6 12:21:46 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6KLkb29713 for linux-xfs-outgoing; Thu, 6 Dec 2001 12:21:46 -0800 Received: from server-15.tower-15.messagelabs.com (mail15.messagelabs.com [63.210.62.243]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6KLfo29688 for ; Thu, 6 Dec 2001 12:21:41 -0800 X-VirusChecked: Checked Received: (qmail 9413 invoked from network); 6 Dec 2001 19:13:54 -0000 Received: from mail.tapeware.com (HELO yt-internet.tapeware.com) (4.21.59.10) by server-15.tower-15.messagelabs.com with SMTP; 6 Dec 2001 19:13:54 -0000 Received: by mail.tapeware.com with Internet Mail Service (5.5.2653.19) id ; Thu, 6 Dec 2001 11:21:32 -0800 Message-ID: From: "Quang Nguyen (Ngo)" To: "'linux-xfs@oss.sgi.com'" Subject: attr_set() Changes ctime Date: Thu, 6 Dec 2001 11:21:30 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Disposition: inline Content-Type: text/plain Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk How do I prevent attr_set() from changing the ctime? I have a feeling that the answer is "you can't". utime() allows to change mtime and atime of a file. Is there a function to change the ctime especially for XFS? Thanks, Quang _____________________________________________________________________ This message has been checked for all known viruses by the MessageLabs Virus Scanning Service. For further information visit http://www.messagelabs.com/stats.asp [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Thu Dec 6 12:28:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6KS3t29991 for linux-xfs-outgoing; Thu, 6 Dec 2001 12:28:03 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6KS0o29968 for ; Thu, 6 Dec 2001 12:28:00 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id LAA04629 for ; Thu, 6 Dec 2001 11:28:02 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id NAA3774209 for ; Thu, 6 Dec 2001 13:26:42 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id NAA92816 for ; Thu, 6 Dec 2001 13:26:42 -0600 (CST) From: Steve Lord Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB6JQBv07579; Thu, 6 Dec 2001 13:26:11 -0600 Message-Id: <200112061926.fB6JQBv07579@jen.americas.sgi.com> Date: Thu, 6 Dec 2001 13:26:11 -0600 Subject: TAKE - fix growfs taking a file across the 1 Tbyte boundary Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Thu Dec 6 11:26:09 PST 2001 Workarea: jen.americas.sgi.com:/src/lord/xfs-linux.2.4 The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107938a linux/fs/xfs/xfs_fsops.c - 1.72 - Add in the 32 bit inode mount flag before re initializing the perag structures in growfs. From owner-linux-xfs@oss.sgi.com Thu Dec 6 12:35:09 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6KZ9830327 for linux-xfs-outgoing; Thu, 6 Dec 2001 12:35:09 -0800 Received: from mxzilla4.xs4all.nl (mxzilla4.xs4all.nl [194.109.6.48]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6KZ5o30305 for ; Thu, 6 Dec 2001 12:35:05 -0800 Received: from xs4.xs4all.nl (knuffie@xs4.xs4all.nl [194.109.6.45]) by mxzilla4.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB6JZ0Ea052745; Thu, 6 Dec 2001 20:35:00 +0100 (CET) Received: from localhost (knuffie@localhost) by xs4.xs4all.nl (8.9.0/8.9.0) with ESMTP id UAA09321; Thu, 6 Dec 2001 20:35:00 +0100 (CET) Date: Thu, 6 Dec 2001 20:35:00 +0100 (CET) From: Seth Mos To: "Drew 'Cheetah!' Maxwell" cc: Jesse Hall , linux-xfs@oss.sgi.com Subject: Re: Linux w/XFS on Athlon systems In-Reply-To: <3C0FB914.51EED80B@clubphoto.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 6 Dec 2001, Drew 'Cheetah!' Maxwell wrote: > Jesse Hall wrote: > -------------------------------------------------------------------------------- > > > > What do you mean by "haven't worked"? > > > > I've been using a Tyan Thunder K7 with dual Athlon MPs with XFS (tracking > > CVS) for a few months with great reliability and stability. > > > By "haven't worked", I mean generating a Kernel Oops, and never being able to > finish bringing the system up to multiuser mode. (It'll die somewhere in the > midst of the interactive menus as it brings up daemons.) Sounds to me like RAM with a lot of holes in it. Get some/new other ones and make sure that the processor is not overheating and such. Cheers From owner-linux-xfs@oss.sgi.com Thu Dec 6 13:20:28 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6LKSF31290 for linux-xfs-outgoing; Thu, 6 Dec 2001 13:20:28 -0800 Received: from roper.uwyo.edu (pmdf@roper.uwyo.edu [129.72.10.8] (may be forged)) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6LKLo31268 for ; Thu, 6 Dec 2001 13:20:21 -0800 Received: from CONVERSION-DAEMON by ROPER.UWYO.EDU (PMDF V5.2-32 #33749) id <0GNX00001V03Q5@ROPER.UWYO.EDU> for linux-xfs@oss.sgi.com; Thu, 6 Dec 2001 13:18:28 -0700 (MST) Received: from asuwlink.uwyo.edu (asuwlink.uwyo.edu [129.72.60.2]) by ROPER.UWYO.EDU (PMDF V5.2-32 #33749) with ESMTP id <0GNX00MCQUYFFT@ROPER.UWYO.EDU> for linux-xfs@oss.sgi.com; Thu, 06 Dec 2001 13:15:51 -0700 (MST) Received: from uwyo.edu (ringram@ross226.uwyo.edu [129.72.51.220]) by asuwlink.uwyo.edu (8.8.8/8.8.7) with ESMTP id NAA13457 for ; Thu, 06 Dec 2001 13:15:50 -0700 (MST) Date: Thu, 06 Dec 2001 13:28:36 +0000 From: ringram@uwyo.edu Subject: Re: Linux w/XFS on Athlon systems To: linux-xfs@oss.sgi.com Reply-to: ringram@uwyo.edu Message-id: <3C0F7284.D452DFAE@uwyo.edu> Organization: UW Math Dept/Institute for Scientific Computation MIME-version: 1.0 X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.13 i686) Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit X-Accept-Language: en References: <3C0FB3D0.B23E6185@clubphoto.com> Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk "Drew 'Cheetah!' Maxwell" wrote: > > Fellow XFS users: > -------------------------------------------------------------------------------- > My company has been working with XFS-Linux on Intel systems for some > time > and has had some difficulties getting a stable Linux w/XFS release to > work > on some of our older K6-2 systems. Instead of butting our heads against > a > case of potentially suspect hardware, we're looking to upgrade to a > group > of Athlon-based systems. In this, I have a few questions: > > 1. What can those of you running Linux w/XFS recommend for motherboards > to > go with an Athlon processor? > (We might as well get hardware we know works by experience.) > 2. Are they any special changes to your kernels used to work correctly > with > an Athlon-based system? > > Our initial attempts at doing an Athlon + Linux + XFS solution haven't > worked. > We've tried the following methods: > 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using > gcc > 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using > kgcc > Pre-patched XFS 1.0.2 kernel source from SGI, compiled using kgcc > > Any assistance is greatly appreciated. > > --Drew 'Cheetah!' Maxwell [cheetah@clubphoto.com] I can almost gaurantee you the problem is not with the Athlon/XFS combination. I have 7 systems here running on Athlon/Duron processors for close to 6 months now with no problems. I also have 2 at home. You need to do some better checking into what is really causing your oops. -- Russel H. Ingram Unix Systems Administrator Institute for Scientific Computation University of Wyoming/Math Dept. Phone: (307)766-6546 E-Mail: ringram@uwyo.edu From owner-linux-xfs@oss.sgi.com Thu Dec 6 14:09:36 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6M9aF32385 for linux-xfs-outgoing; Thu, 6 Dec 2001 14:09:36 -0800 Received: from anime.net (root@anime.net [63.172.78.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6M9Xo32363 for ; Thu, 6 Dec 2001 14:09:33 -0800 Received: from localhost (goemon@localhost) by anime.net (8.11.6/8.9.3) with ESMTP id fB6L99A30689; Thu, 6 Dec 2001 13:09:09 -0800 Date: Thu, 6 Dec 2001 13:09:09 -0800 (PST) From: Dan Hollis To: Stefan Smietanowski cc: Simon Matter , Steve Lord , linux xfs ml Subject: Re: How to fsck read-only-mounted root filesystem? In-Reply-To: <3C0F7F65.7040809@stesmi.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 6 Dec 2001, Stefan Smietanowski wrote: > Dan Hollis wrote: > > Suprised in the respect that ext2 and reiserfs have no such limitation, I > > was expecting xfs to be at least as functional and mature. > It's just designed differently. > If You go buy a ferrari and go whine that you can't fit that piano in > the trunk, then you have to think what's important for you. Fitting the > piano or driving fast. If repair-on-mounted-filesystem is what you're > after then you use a filesystem that supports it. If you're after a > good, fast and stable (and mature) filesystem you stick to xfs. The > choice is yours. If I had known xfs couldn't repair read-only-mounted filesystems, I would never have installed it. I suspect other people will make the same decision, if they are actually aware of this limitation. This xfs limitation is stated NOWHERE. This warning should go into the FAQ. It would have saved a lot of headaches. :-( -Dan -- [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Thu Dec 6 14:11:19 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6MBJp32552 for linux-xfs-outgoing; Thu, 6 Dec 2001 14:11:19 -0800 Received: from anime.net (root@anime.net [63.172.78.150]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6MBHo32530 for ; Thu, 6 Dec 2001 14:11:17 -0800 Received: from localhost (goemon@localhost) by anime.net (8.11.6/8.9.3) with ESMTP id fB6LBAW30727; Thu, 6 Dec 2001 13:11:10 -0800 Date: Thu, 6 Dec 2001 13:11:10 -0800 (PST) From: Dan Hollis To: Seth Mos cc: Xianglong Yuan , Subject: Re: Follow up -- Re: Files on XFS not safe?! In-Reply-To: <4.3.2.7.2.20011206084345.02e5c618@pop.xs4all.nl> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 6 Dec 2001, Seth Mos wrote: > Tux2 seems to be a really good idea for static content systems but I am > afraid that databases will not like phase trees. Why would they care? I can't see that it would make a single difference at all. -Dan -- [-] Omae no subete no kichi wa ore no mono da. [-] From owner-linux-xfs@oss.sgi.com Thu Dec 6 14:29:18 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6MTIs00618 for linux-xfs-outgoing; Thu, 6 Dec 2001 14:29:18 -0800 Received: from clubphoto.com (mail-gw.clubphoto.com [64.124.34.66]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6MTEo00595 for ; Thu, 6 Dec 2001 14:29:14 -0800 Received: from [192.168.168.125] ([192.168.168.125]) by clubphoto.com (8.9.3/8.9.3) with ESMTP id NAA10320; Thu, 6 Dec 2001 13:29:03 -0800 User-Agent: Microsoft-Entourage/10.0.0.1309 Date: Thu, 06 Dec 2001 13:29:15 -0800 Subject: Re: Linux w/XFS on Athlon systems From: "Gabe E. Nydick" To: Seth Mos , "Drew 'Cheetah!' Maxwell" CC: Jesse Hall , Message-ID: In-Reply-To: Mime-version: 1.0 Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Seth, The system we have works fine with 1000 simultaneous bonnie++ processes running under a PIII kernel. Do you think it could still be the memory? Thanks, Gabe On 12/6/01 11:35 AM, "Seth Mos" wrote: > On Thu, 6 Dec 2001, Drew 'Cheetah!' Maxwell wrote: > >> Jesse Hall wrote: >> ----------------------------------------------------------------------------- >> --- >>> >>> What do you mean by "haven't worked"? >>> >>> I've been using a Tyan Thunder K7 with dual Athlon MPs with XFS (tracking >>> CVS) for a few months with great reliability and stability. >>> >> By "haven't worked", I mean generating a Kernel Oops, and never being able to >> finish bringing the system up to multiuser mode. (It'll die somewhere in the >> midst of the interactive menus as it brings up daemons.) > > Sounds to me like RAM with a lot of holes in it. > > Get some/new other ones and make sure that the processor is not > overheating and such. > > Cheers > From owner-linux-xfs@oss.sgi.com Thu Dec 6 14:30:57 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6MUvb00814 for linux-xfs-outgoing; Thu, 6 Dec 2001 14:30:57 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6MUqo00788 for ; Thu, 6 Dec 2001 14:30:52 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id NAA06164 for ; Thu, 6 Dec 2001 13:30:38 -0800 (PST) mail_from (roehrich@sgi.com) Received: from thistle-e185.americas.sgi.com (thistle-e185.americas.sgi.com [128.162.185.204]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id PAA3675758 for ; Thu, 6 Dec 2001 15:29:34 -0600 (CST) Received: from clink.americas.sgi.com (clink-eth.americas.sgi.com [128.162.2.8]) by thistle-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id PAA00945 for ; Thu, 6 Dec 2001 15:29:34 -0600 (CST) Received: (from roehrich@localhost) by clink.americas.sgi.com (SGI-8.9.3/8.9.3/erikj-IRIX) id PAA63823 for linux-xfs@oss.sgi.com; Thu, 6 Dec 2001 15:29:33 -0600 (CST) Date: Thu, 6 Dec 2001 15:29:33 -0600 (CST) From: Dean Roehrich Message-Id: <200112062129.PAA63823@clink.americas.sgi.com> To: linux-xfs@oss.sgi.com Subject: TAKE - rip out xfs_dm_fcntl, bypass it Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Thu Dec 6 13:29:13 PST 2001 Workarea: clink-eth.americas.sgi.com:/data/clink/a67/roehrich/2.4.x-xfs The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107944a linux/fs/xfs/xfs_dmapi.h - 1.19 - Remove prototype for xfs_dm_fcntl(). Remove prototype for xfs_set_dmattrs, it was used by xfs_dm_fcntl. Update dm_fcntl_* data structure names. linux/fs/xfs/xfs_dmapi.c - 1.45 - Rip out xfs_dm_fcntl(). xfs_dm_mapevent() and xfs_dm_get_fsys_vector() should bypass the data structure for xfs_dm_fcntl() since we're no longer using it. linux/fs/xfs/xfsdmapistubs.c - 1.10 - Rip out xfs_dm_fcntl. linux/fs/xfs/linux/xfs_file.c - 1.55 - linvfs_dmapi_map_event() should bypass the data structure for xfs_dm_fcntl() since we're no longer using it. linux/fs/xfs/linux/xfs_ioctl.c - 1.51 - xfs_fssetdm_by_handle() was the last user of xfs_dm_fcntl(), and the only effect was to get a redundant check on CAP_MKNOD. Bypass it, and go straight to xfs_set_dmattrs(). linux/include/linux/dmapi_kern.h - 1.7 - Update comments. Rip out structures and symbols that are no longer used. linux/fs/xfs/linux/xfs_vfs.h - 1.8 - Bypass xfs_dm_fcntl(), go straight to xfs_dm_get_fsys_vector(). cmd/dmapi/include/dmapi_kern.h - 1.2 - match linux/include/linux/dmapi_kern.h linux/fs/xfs_dmapi/dmapi_mountinfo.c - 1.3 - dm_query_fsys_for_vector() should bypass the data structure for xfs_dm_fcntl() since we're no longer using it. From owner-linux-xfs@oss.sgi.com Thu Dec 6 15:00:59 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6N0xc01706 for linux-xfs-outgoing; Thu, 6 Dec 2001 15:00:59 -0800 Received: from mail.get2chip.com ([64.169.83.2]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6N0no01679 for ; Thu, 6 Dec 2001 15:00:49 -0800 Message-ID: <3C0FE687.C1238FB6@get2chip.com> Date: Thu, 06 Dec 2001 13:43:35 -0800 From: ccroswhite@get2chip.com Reply-To: ccroswhite@get2chip.com MIME-Version: 1.0 To: linux-xfs@oss.sgi.com Subject: UW Math Dept/Institute for Scientific Computation Content-Type: multipart/mixed; boundary="------------B8B072FB390C6F4975C146FE" Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk This is a multi-part message in MIME format. --------------B8B072FB390C6F4975C146FE Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit "Drew 'Cheetah!' Maxwell" wrote: > > Fellow XFS users: > -------------------------------------------------------------------------------- > My company has been working with XFS-Linux on Intel systems for some > time > and has had some difficulties getting a stable Linux w/XFS release to > work > on some of our older K6-2 systems. Instead of butting our heads against > a > case of potentially suspect hardware, we're looking to upgrade to a > group > of Athlon-based systems. In this, I have a few questions: > > 1. What can those of you running Linux w/XFS recommend for motherboards > to > go with an Athlon processor? > (We might as well get hardware we know works by experience.) > 2. Are they any special changes to your kernels used to work correctly > with > an Athlon-based system? > > Our initial attempts at doing an Athlon + Linux + XFS solution haven't > worked. > We've tried the following methods: > 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using > gcc > 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using > kgcc > Pre-patched XFS 1.0.2 kernel source from SGI, compiled using kgcc > > Any assistance is greatly appreciated. We have been running a couple dozen Tyan S2460 w/2x Athlon 1800+ (yep, not the MP's) with XFS, from the revamped SGI installers+RH7.2, for close to a month. We have never had a problem, no hicups, nada nothing! Seems perhaps it might be hardware related?!?! Chris Croswhite Get2Chip, Inc. San Jose, CA USA 95131 001.408.501.9525 --------------B8B072FB390C6F4975C146FE Content-Type: text/x-vcard; charset=us-ascii; name="ccroswhite.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Content-Disposition: attachment; filename="ccroswhite.vcf" begin:vcard n:Croswhite;Chris tel;work:001.408.501.9600 - 9525 x-mozilla-html:FALSE url:www.get2chip.com org:Get2Chip, Inc.;IT/IS version:2.1 email;internet:ccroswhite@get2chip.com adr;quoted-printable:;;2107 North First Street=0D=0ASuite 350;San Jose;CA;95131;USA x-mozilla-cpt:;-7008 fn:Chris Croswhite end:vcard --------------B8B072FB390C6F4975C146FE-- From owner-linux-xfs@oss.sgi.com Thu Dec 6 15:11:40 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB6NBeq02033 for linux-xfs-outgoing; Thu, 6 Dec 2001 15:11:40 -0800 Received: from sunfish.linuxis.net (sunfish.linuxis.net [64.71.162.66]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB6NBbo02011 for ; Thu, 6 Dec 2001 15:11:37 -0800 Received: (qmail 19952 invoked by uid 1000); 6 Dec 2001 22:06:59 -0000 Date: Thu, 6 Dec 2001 14:06:57 -0800 To: linux-xfs@oss.sgi.com Subject: Re: Linux w/XFS on Athlon systems Message-ID: <20011206140657.O13813@flounder.net> Mail-Followup-To: linux-xfs@oss.sgi.com References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.3.21i X-Delivery-Agent: TMDA v0.42/Python 2.1.1 (sunos5) From: "Adam McKenna" Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, Dec 06, 2001 at 01:29:15PM -0800, Gabe E. Nydick wrote: > Seth, > > The system we have works fine with 1000 simultaneous bonnie++ processes > running under a PIII kernel. Do you think it could still be the memory? What is the IDE chipset? XFS is known to exacerbate problems with some chipsets due to its heavier disk use. --Adam From owner-linux-xfs@oss.sgi.com Thu Dec 6 16:00:59 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB700xY03200 for linux-xfs-outgoing; Thu, 6 Dec 2001 16:00:59 -0800 Received: from smtpzilla2.xs4all.nl (smtpzilla2.xs4all.nl [194.109.127.138]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB700qo03178 for ; Thu, 6 Dec 2001 16:00:53 -0800 Received: from auto-nb1.xs4all.nl ([212.58.167.191]) by smtpzilla2.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB6N0gm5069243; Fri, 7 Dec 2001 00:00:47 +0100 (CET) Message-Id: <4.3.2.7.2.20011206234949.023196a8@pop.xs4all.nl> X-Sender: knuffie@pop.xs4all.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Thu, 06 Dec 2001 23:57:21 +0100 To: "Gabe E. Nydick" , "Drew 'Cheetah!' Maxwell" From: Seth Mos Subject: Re: Linux w/XFS on Athlon systems Cc: Jesse Hall , In-Reply-To: References: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 13:29 6-12-2001 -0800, Gabe E. Nydick wrote: >Seth, > > The system we have works fine with 1000 simultaneous bonnie++ processes >running under a PIII kernel. Do you think it could still be the memory? I had one system with 64 MB of ram of which about 1 was faulty. Sometimes it booted. It died at various places. corrupted some files crashed some processes, some stuck in D state, hanging kupdate, oopses, panics. I had seen about every error known to man in about 2 hours untill I removed one of the dims and everything was well. That was 4 months ago. The machine is still running to this date and has never crashed _once_ since that time or even give any form of trouble. Bonnie excercises disk. The Bonnie proces stays in RAM and your disk cache might wander right over the bad piece of ram without crashing the kernel. You enable the Athlon option, it reorganizes some ways of reading and writing to the ram to speed it up and suddenly the kernel code is trying to use the faulty ram. Seems plausible to me. Did you enable mtrr by any chance? I could make a Eicon Diva server card stop working by enabling or disabling mtrr. Because of remapping some buffers and PCI ports the driver was not the same spot where the driver thought it should have been. -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Thu Dec 6 16:11:00 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB70B0i03481 for linux-xfs-outgoing; Thu, 6 Dec 2001 16:11:00 -0800 Received: from smtpzilla2.xs4all.nl (smtpzilla2.xs4all.nl [194.109.127.138]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB70Aso03455 for ; Thu, 6 Dec 2001 16:10:55 -0800 Received: from auto-nb1.xs4all.nl ([212.58.167.191]) by smtpzilla2.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB6NAjm5070409; Fri, 7 Dec 2001 00:10:49 +0100 (CET) Message-Id: <4.3.2.7.2.20011206235736.034fd528@pop.xs4all.nl> X-Sender: knuffie@pop.xs4all.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Fri, 07 Dec 2001 00:07:25 +0100 To: Dan Hollis From: Seth Mos Subject: Re: Follow up -- Re: Files on XFS not safe?! Cc: Xianglong Yuan , In-Reply-To: References: <4.3.2.7.2.20011206084345.02e5c618@pop.xs4all.nl> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 13:11 6-12-2001 -0800, Dan Hollis wrote: >On Thu, 6 Dec 2001, Seth Mos wrote: > > Tux2 seems to be a really good idea for static content systems but I am > > afraid that databases will not like phase trees. > >Why would they care? Databases work synchronously most of the time for the data but they might use a buffer for storing intermediate before they are actually written out. This happens with transactions. they store all the data which needs to be committed in the end to actual database or be discarded when it doesn't procede. >I can't see that it would make a single difference at all. It probably won't, I think my brain just switched on (the horror). I was mostly thinking of it as something that writes in batches which should be faster but the actual latency might get higher. But I might have interpreted the scheme too course. If it is something that runs every 1/20 second you would get a continous stream of IO but I am still wondering about when multiple file being touched at once (or multiple databases if you like) and meaning that a read must wait before the tree has been commited to disk which might take long if you toch a lot of files at once, before returning the read call. These are a lot of "mights" and still havn't read the article. Include standard idiot disclaimer here. Cheers -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Thu Dec 6 16:11:35 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB70BZ903623 for linux-xfs-outgoing; Thu, 6 Dec 2001 16:11:35 -0800 Received: from dream.lapseofthought.com (adsl-66-124-80-202.dsl.snfc21.pacbell.net [66.124.80.202]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB70BVo03601 for ; Thu, 6 Dec 2001 16:11:31 -0800 Received: (from drich@localhost) by dream.lapseofthought.com (8.11.6/8.11.6) id fB6NBUK02221; Thu, 6 Dec 2001 15:11:30 -0800 Date: Thu, 6 Dec 2001 15:11:30 -0800 (PST) From: Dan Rich X-X-Sender: To: Subject: XFS (Linux) and locking problems Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk I am seeing odd problems on a Linux XFS filesystem that appear to be caused by a locking issue. I have had several files dissapear and I believe it is due to two processes (mail filters) opening the file at the same time and steping on each other's toes. Have you seen anything like this on other systems? I'm seeing this with the XFS 1.0.1 release on the 2.4.9 kernel. I haven't tested the 1.0.2 release yet due to problems with the 2.4.14 kernel and the 1.0.2 build on my system. Also, is there a timetable for releaseing XFS patches for 2.4.16? Thanks! -- Dan Rich | http://www.employees.org/~drich/ | "Step up to red alert!" "Are you sure, sir? | It means changing the bulb in the sign..." | - Red Dwarf (BBC) From owner-linux-xfs@oss.sgi.com Thu Dec 6 16:16:45 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB70GjL03856 for linux-xfs-outgoing; Thu, 6 Dec 2001 16:16:45 -0800 Received: from rj.sgi.com (rj.sgi.com [204.94.215.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB70Gfo03834 for ; Thu, 6 Dec 2001 16:16:41 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by rj.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB6NGZY00603 for ; Thu, 6 Dec 2001 15:16:35 -0800 Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id RAA3777397; Thu, 6 Dec 2001 17:15:20 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id RAA74919; Thu, 6 Dec 2001 17:15:19 -0600 (CST) Subject: Re: XFS (Linux) and locking problems From: Eric Sandeen To: Dan Rich Cc: linux-xfs@oss.sgi.com In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 06 Dec 2001 17:15:18 -0600 Message-Id: <1007680518.13468.22.camel@stout.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 2001-12-06 at 17:11, Dan Rich wrote: > I am seeing odd problems on a Linux XFS filesystem that appear to be > caused by a locking issue. I have had several files dissapear and I > believe it is due to two processes (mail filters) opening the file at the > same time and steping on each other's toes. Have you seen anything like > this on other systems? I'm seeing this with the XFS 1.0.1 release on > the 2.4.9 kernel. Hm, there was no 1.0.1 release for the 2.4.9 kernel... Disappear... as in, no longer in the directory listing? Or contents disappear? > I haven't tested the 1.0.2 release yet due to problems with the 2.4.14 > kernel and the 1.0.2 build on my system. > > Also, is there a timetable for releaseing XFS patches for 2.4.16? They're available in the patches/ dir as of last night or so. -Eric -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Thu Dec 6 16:17:00 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB70H0703937 for linux-xfs-outgoing; Thu, 6 Dec 2001 16:17:00 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB70Gko03866 for ; Thu, 6 Dec 2001 16:16:46 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via SMTP id AAA1616605 for ; Fri, 7 Dec 2001 00:16:43 +0100 (CET) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id KAA19905; Fri, 7 Dec 2001 10:15:21 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id KAA53637; Fri, 7 Dec 2001 10:15:18 +1100 (AEDT) Date: Fri, 7 Dec 2001 10:15:17 +1100 From: Nathan Scott To: Daniel Phillips , Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [PATCH] Revised extended attributes interface Message-ID: <20011207101517.B46546@wobbly.melbourne.sgi.com> References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <20011206164131.F50483@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from phillips@bonn-fries.net on Thu, Dec 06, 2001 at 04:25:41PM +0100 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi Daniel, On Thu, Dec 06, 2001 at 04:25:41PM +0100, Daniel Phillips wrote: > On December 6, 2001 06:41 am, Nathan Scott wrote: > > > > I guess it ultimately comes down to simplicity. > > Right now we have two namespaces, user and system. Andreas and the security folk have long been investigating "trusted" and "owner" namespaces too. See Andreas' web pages for more discussion on those. I see no reason to impose such arbitrary restrictions as you seem to be suggesting - if a filesystem does not want to or cannot implement some particular namespace, then it shouldn't be forced to. eg. there is no reason why the user namespace has to be implemented - Andreas patches for ext2/ext3 go so far as to make this namespace compile-time conditional (and his userspace tools just continue to work - its really quite a nice design, IMO). > That's one bit of > information, and the proposal is to represent it with 5-7 bytes, passing it > on every call, and decoding it with a memcmp or similar. This is just extra > fluff as far as I can see, I'd be interested in seeing exactly how you'd like to see the interfaces changed - could you put forward some APIs? It sounds like you're suggesting a separate integer namespace parameter to each syscall/vfs interface? I think you'll find this solves none of the problems you're describing, and makes every operation more complex, and more difficult. Worse, its alot more open to abuse in the way the old command parameter was than namespace-prefixed names are! (there would probably be some free high-order bits in there where I could sneak new functionality in, right?) But perhaps I'm misunderstanding what you're suggesting - I should wait to see your patch. > and provides every bit as much opportunity for > implementing a private API as the original cmd parameter did, by encoding > whatever one pleases before the dot. This is just not true - the API does not change at all if a new namespace was needed for some reason. Restricting namespaces by using a binary namespace also doesn't help at all - the namespace becomes obfuscated (strings are easier to grok than bitfields), divorced from the name (which it clarifies, making life difficult for the userspace tools) and doesn't even solve the "problem" - someone could use new bits just as easily as use new namespace strings. > > (unless you go for a binary namespace representation, and that's a > > real can of worms). > > I'm suggesting we take a look at that. > Andreas and I did have such an implementation, but we ditched it. The CVS revision history of cmd/attr2/{set,get}fattr/*.c in the XFS tree show the progression of user<->kernel interfaces which I tried while Andreas and I were nutting out a clean solution that we both could use. Thar be dragons thar. Big hairy ones. > > [extended attributes on symlinks] > > OK, well it looks like you're going a little overboard here in dividing out > the functionality. What you're talking about is 'follow symlink or not', > right? That really does sound to me as though it's naturally expressed with > a flag bit. I really don't see a compelling reason to go beyond 8 syscalls: > > get, fget, set, fset, del, fdel, list, flist > I'm not too fussed - the second draft patch I sent out did exactly as you describe, in an attempt to cut down on syscalls. This again meant adding a "flags" field to each operation. We also have stat/ lstat/fstat and chown/lchown/fchown - I was trying to be consistent with those, and I still think that is the right thing to do. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Thu Dec 6 16:23:11 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB70NB504181 for linux-xfs-outgoing; Thu, 6 Dec 2001 16:23:11 -0800 Received: from zok.sgi.com (zok.SGI.COM [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB70N8o04159 for ; Thu, 6 Dec 2001 16:23:08 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB6NN2A02840 for ; Thu, 6 Dec 2001 15:23:02 -0800 Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id RAA3758718 for ; Thu, 6 Dec 2001 17:21:46 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id RAA74819 for ; Thu, 6 Dec 2001 17:21:46 -0600 (CST) From: Steve Lord Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB6NLDR18270; Thu, 6 Dec 2001 17:21:13 -0600 Message-Id: <200112062321.fB6NLDR18270@jen.americas.sgi.com> Date: Thu, 6 Dec 2001 17:21:13 -0600 Subject: TAKE - fix ext3 run-bash-shared-mappings.sh test on xfs Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk This test used to deadlock. Date: Thu Dec 6 15:20:41 PST 2001 Workarea: jen.americas.sgi.com:/src/lord/xfs-linux.2.4 The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:107953a linux/fs/pagebuf/page_buf_io.c - 1.101 - Fix xfs when running the ext3 run-bash-shared-mappings.sh test - deal with the case of doing read/write file I/O to/from a mmapped region in the same file. From owner-linux-xfs@oss.sgi.com Thu Dec 6 16:25:44 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB70Pi504348 for linux-xfs-outgoing; Thu, 6 Dec 2001 16:25:44 -0800 Received: from zok.sgi.com (zok.SGI.COM [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB70Pfo04304 for ; Thu, 6 Dec 2001 16:25:41 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB6NPZA03007 for ; Thu, 6 Dec 2001 15:25:35 -0800 Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id RAA3788112 for ; Thu, 6 Dec 2001 17:24:19 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id RAA99495 for ; Thu, 6 Dec 2001 17:24:18 -0600 (CST) Subject: 2.5 CVS available From: Eric Sandeen To: linux-xfs@oss.sgi.com Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 06 Dec 2001 17:24:17 -0600 Message-Id: <1007681057.13482.24.camel@stout.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Yes, that's right, act now to be the first on your block to jeopardize your data, and run XFS under Linux 2.5.1-pre5! Send no money now, simply: $ export CVSROOT=':pserver:cvs@oss.sgi.com:/cvs' $ cvs login (pw: cvs) $ cvs -z3 checkout linux-2.5-xfs Remember, 2.5 is Development with a capital "D" so don't use this for anything that _matters_... -Eric -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Thu Dec 6 16:25:48 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB70PmP04404 for linux-xfs-outgoing; Thu, 6 Dec 2001 16:25:48 -0800 Received: from rj.sgi.com (rj.sgi.com [204.94.215.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB70Pfo04305 for ; Thu, 6 Dec 2001 16:25:41 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by rj.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with SMTP id fB6NPWY01186 for ; Thu, 6 Dec 2001 15:25:32 -0800 Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id KAA19959; Fri, 7 Dec 2001 10:24:14 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id KAA53534; Fri, 7 Dec 2001 10:24:13 +1100 (AEDT) Date: Fri, 7 Dec 2001 10:24:12 +1100 From: Nathan Scott To: vkuznet Cc: linux-xfs@oss.sgi.com Subject: Re: XFS and quota problem Message-ID: <20011207102411.C46546@wobbly.melbourne.sgi.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from vkuznet@fnal.gov on Thu, Dec 06, 2001 at 11:15:15AM -0600 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi Valentin, On Thu, Dec 06, 2001 at 11:15:15AM -0600, vkuznet wrote: > I have a problem with quota command on XFS partitions. > We have a cluster of ~150 nodes with XFS fs and after > upgrading to 2.4.14 kernel our quota -v refuse to > work. The source RPM for quota package from SGI (found on > rpmfind.net) compiled successfully but show 0 as your That is actually a fairly old version of the code now, you'll see we no longer ship that in recent releases. You can use the quota code that comes with Redhat now (since 7.2 IIRC), or download the source from quota-tools project on sourceforge. I have a vague recollection of reports of this problem being fixed by an upgrade - if not, let me know what happens with new tools & we'll take it from there. cheers. > quota limit. Example: > > > vkuznet@vk-clued0(12:40:37)> usr/bin/quota -v > > Disk quotas for user vkuznet (uid 1212): > > Filesystem blocks quota limit grace files quota limit > > grace > > /d02ka.fnal.gov:/desktop/home5/vkuznet > > 0 0 0 0 0 0 > > /plum-clued0:/disk2/fnal/ups > > 0 0 0 0 0 0 > > /plum-clued0.fnal.gov:/disk2/local > > 0 0 0 0 0 0 > > We do have a variety of disks (local, cross mounted from Linux and > from SGI file server) and none of them show valid values for quotas. > > Can anyone explain the problem and/or provide or point me to valid patch. > > I'm not on XFS maling list (sorry too many mails) and will appreciate > if you replay to my Email directly. > > Thanks, > Valentin. > -- Nathan From owner-linux-xfs@oss.sgi.com Thu Dec 6 16:41:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB70f3Z04737 for linux-xfs-outgoing; Thu, 6 Dec 2001 16:41:03 -0800 Received: from Cantor.suse.de (ns.suse.de [213.95.15.193]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB70f0o04715 for ; Thu, 6 Dec 2001 16:41:00 -0800 Received: from Hermes.suse.de (Hermes.suse.de [213.95.15.136]) by Cantor.suse.de (Postfix) with ESMTP id 12B961E52B; Fri, 7 Dec 2001 00:40:54 +0100 (MET) Date: Fri, 7 Dec 2001 00:40:53 +0100 From: Andi Kleen To: Eric Sandeen Cc: linux-xfs@oss.sgi.com Subject: Re: 2.5 CVS available Message-ID: <20011207004053.A22133@wotan.suse.de> References: <1007681057.13482.24.camel@stout.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1007681057.13482.24.camel@stout.americas.sgi.com> User-Agent: Mutt/1.3.22.1i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, Dec 06, 2001 at 05:24:17PM -0600, Eric Sandeen wrote: > Yes, that's right, act now to be the first on your block to jeopardize > your data, and run XFS under Linux 2.5.1-pre5! > > Send no money now, simply: > > $ export CVSROOT=':pserver:cvs@oss.sgi.com:/cvs' > $ cvs login (pw: cvs) > $ cvs -z3 checkout linux-2.5-xfs > > Remember, 2.5 is Development with a capital "D" so don't use this for > anything that _matters_... Could you give a short summary on the changes? Does it use the new BIO interface yet for scatter-gather similar to the old kiovec path ? -Andi From owner-linux-xfs@oss.sgi.com Thu Dec 6 16:50:42 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB70og505864 for linux-xfs-outgoing; Thu, 6 Dec 2001 16:50:42 -0800 Received: from lips.thebarn.com (lips.borg.umn.edu [160.94.232.50]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB70oco05842 for ; Thu, 6 Dec 2001 16:50:38 -0800 Received: from scare.vieo.com ([63.231.179.33]) by lips.thebarn.com (8.12.0/8.12.0) with ESMTP id fB6NnaML036023 for ; Thu, 6 Dec 2001 17:49:37 -0600 (CST) Subject: Re: 2.5 CVS available From: Russell Cattelan Cc: linux-xfs@oss.sgi.com In-Reply-To: <1007681057.13482.24.camel@stout.americas.sgi.com> References: <1007681057.13482.24.camel@stout.americas.sgi.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 06 Dec 2001 17:43:00 -0600 Message-Id: <1007682181.2748.12.camel@scare> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 2001-12-06 at 17:24, Eric Sandeen wrote: > Yes, that's right, act now to be the first on your block to jeopardize > your data, and run XFS under Linux 2.5.1-pre5! > > Send no money now, simply: > > $ export CVSROOT=':pserver:cvs@oss.sgi.com:/cvs' > $ cvs login (pw: cvs) > $ cvs -z3 checkout linux-2.5-xfs And for all you fast update addicts *default host=ftp.thebarn.com *default base=. *default release=cvs *default delete use-rel-suffix *default prefix=/tmp/cvsupit *default compress linux2.5-xfs -Russell From owner-linux-xfs@oss.sgi.com Thu Dec 6 17:37:16 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB71bG507825 for linux-xfs-outgoing; Thu, 6 Dec 2001 17:37:16 -0800 Received: from relay-3v.club-internet.fr (relay-3v.club-internet.fr [194.158.96.114]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB71bAo07802 for ; Thu, 6 Dec 2001 17:37:11 -0800 Received: from club-internet.fr (lns01v-2-125.w.club-internet.fr [213.44.241.125]) by relay-3v.club-internet.fr (Postfix) with ESMTP id 5D82116E1; Fri, 7 Dec 2001 01:37:05 +0100 (CET) Message-ID: <3C101038.4A4CA259@club-internet.fr> Date: Fri, 07 Dec 2001 01:41:28 +0100 From: Jean Francois Martinez X-Mailer: Mozilla 4.78 [en] (X11; U; Linux 2.4.9-13SGI_XFS_PR1 i586) X-Accept-Language: en MIME-Version: 1.0 To: Seth Mos , linux-xfs@oss.sgi.com Subject: Re: Linux w/XFS on Athlon systems References: <4.3.2.7.2.20011206234949.023196a8@pop.xs4all.nl> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Seth Mos wrote: > > At 13:29 6-12-2001 -0800, Gabe E. Nydick wrote: > >Seth, > > > > The system we have works fine with 1000 simultaneous bonnie++ processes > >running under a PIII kernel. Do you think it could still be the memory? > > I had one system with 64 MB of ram of which about 1 was faulty. Sometimes > it booted. It died at various places. corrupted some files crashed some > processes, some stuck in D state, hanging kupdate, oopses, panics. I had > seen about every error known to man in about 2 hours untill I removed one > of the dims and everything was well. > > That was 4 months ago. The machine is still running to this date and has > never crashed _once_ since that time or even give any form of trouble. > > Bonnie excercises disk. The Bonnie proces stays in RAM and your disk cache > might wander right over the bad piece of ram without crashing the kernel. > You enable the Athlon option, it reorganizes some ways of reading and > writing to the ram to speed it up and suddenly the kernel code is trying to > use the faulty ram. Seems plausible to me. > > Did you enable mtrr by any chance? I could make a Eicon Diva server card > stop working by enabling or disabling mtrr. Because of remapping some > buffers and PCI ports the driver was not the same spot where the driver > thought it should have been. > Xfree 4.x enable mtrr by default if kernel has been compiled with mtrr support.  -- Jean Francois Martinez Project Independence http://independence.seul.org Because Linux should be for everyone From owner-linux-xfs@oss.sgi.com Thu Dec 6 17:38:53 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB71cro07978 for linux-xfs-outgoing; Thu, 6 Dec 2001 17:38:53 -0800 Received: from zok.sgi.com (zok.SGI.COM [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB71cpo07956 for ; Thu, 6 Dec 2001 17:38:51 -0800 Received: from relay1.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB70cjA07180 for ; Thu, 6 Dec 2001 16:38:45 -0800 Received: from chuckle.americas.sgi.com (sandeen@chuckle.americas.sgi.com [128.162.211.44]) by relay1.corp.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id QAA03103; Thu, 6 Dec 2001 16:38:13 -0800 (PST) Date: Thu, 6 Dec 2001 18:38:13 -0600 (CST) From: Eric Sandeen X-X-Sender: To: Andi Kleen cc: Subject: Re: 2.5 CVS available In-Reply-To: <20011207004053.A22133@wotan.suse.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk I'll let Steve answer, since he did the work. :) -Eric On Fri, 7 Dec 2001, Andi Kleen wrote: > Could you give a short summary on the changes? Does it use the new BIO interface yet for > scatter-gather similar to the old kiovec path ? From owner-linux-xfs@oss.sgi.com Thu Dec 6 17:53:15 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB71rFC08325 for linux-xfs-outgoing; Thu, 6 Dec 2001 17:53:15 -0800 Received: from k-7.stesmi.com (IDENT:root@as4-1-7.has.s.bonet.se [217.215.31.238]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB71rBo08302 for ; Thu, 6 Dec 2001 17:53:12 -0800 Received: from stesmi.com (voyager.stesmi.com [192.168.1.11]) by k-7.stesmi.com (8.11.2/8.8.7) with ESMTP id fB70qSG12981; Fri, 7 Dec 2001 01:52:28 +0100 Message-ID: <3C10134F.5060906@stesmi.com> Date: Fri, 07 Dec 2001 01:54:39 +0100 From: Stefan Smietanowski User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.6) Gecko/20011120 X-Accept-Language: en-us MIME-Version: 1.0 To: Eric Sandeen CC: linux-xfs@oss.sgi.com Subject: Re: 2.5 CVS available References: <1007681057.13482.24.camel@stout.americas.sgi.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi. > Yes, that's right, act now to be the first on your block to jeopardize > your data, and run XFS under Linux 2.5.1-pre5! > > Send no money now, simply: > > $ export CVSROOT=':pserver:cvs@oss.sgi.com:/cvs' > $ cvs login (pw: cvs) > $ cvs -z3 checkout linux-2.5-xfs > > Remember, 2.5 is Development with a capital "D" so don't use this for > anything that _matters_... Ok, any diff in the XFS part of the kernel from the 2.4.x tree? // Stefan From owner-linux-xfs@oss.sgi.com Thu Dec 6 18:23:17 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB72NHc08808 for linux-xfs-outgoing; Thu, 6 Dec 2001 18:23:17 -0800 Received: from server-3.tower-15.messagelabs.com (mail15.messagelabs.com [63.210.62.243]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB72NAo08781 for ; Thu, 6 Dec 2001 18:23:11 -0800 X-VirusChecked: Checked Received: (qmail 8812 invoked from network); 7 Dec 2001 01:12:39 -0000 Received: from mail.tapeware.com (HELO yt-internet.tapeware.com) (4.21.59.10) by server-3.tower-15.messagelabs.com with SMTP; 7 Dec 2001 01:12:39 -0000 Received: by mail.tapeware.com with Internet Mail Service (5.5.2653.19) id ; Thu, 6 Dec 2001 17:18:26 -0800 Message-ID: From: "Quang Nguyen (Ngo)" To: "Quang Nguyen (Ngo)" , "''linux-xfs@oss.sgi.com' '" Subject: RE: attr_set() Changes ctime Date: Thu, 6 Dec 2001 17:18:24 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Disposition: inline Content-Type: text/plain Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Does anyone have an answer to this question? ;-) Thanks, Quang -----Original Message----- From: Quang Nguyen (Ngo) To: 'linux-xfs@oss.sgi.com' Sent: 12/6/01 11:21 AM Subject: attr_set() Changes ctime How do I prevent attr_set() from changing the ctime? I have a feeling that the answer is "you can't". utime() allows to change mtime and atime of a file. Is there a function to change the ctime especially for XFS? Thanks, Quang _____________________________________________________________________ This message has been checked for all known viruses by the MessageLabs Virus Scanning Service. For further information visit http://www.messagelabs.com/stats.asp [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Thu Dec 6 18:39:59 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB72dx609113 for linux-xfs-outgoing; Thu, 6 Dec 2001 18:39:59 -0800 Received: from johnson.mail.mindspring.net (johnson.mail.mindspring.net [207.69.200.177]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB72dpo09091 for ; Thu, 6 Dec 2001 18:39:51 -0800 Received: from user-uini6gv.dsl.mindspring.com ([165.121.26.31] helo=waltsathlon.localhost.net) by johnson.mail.mindspring.net with esmtp (Exim 3.33 #1) id 16C9zR-0003AJ-00 for linux-xfs@oss.sgi.com; Thu, 06 Dec 2001 20:39:49 -0500 Received: from mindspring.com (localhost.localdomain [127.0.0.1]) by waltsathlon.localhost.net (Postfix) with ESMTP id 198BFD03FE11; Thu, 6 Dec 2001 17:30:20 -0800 (PST) Message-ID: <3C101BAB.1060901@mindspring.com> Date: Thu, 06 Dec 2001 17:30:19 -0800 From: Walt H User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.6+) Gecko/20011205 X-Accept-Language: en-us MIME-Version: 1.0 To: "Drew 'Cheetah!' Maxwell" Cc: linux-xfs@oss.sgi.com Subject: Re: Linux w/XFS on Athlon systems References: <3C0FB3D0.B23E6185@clubphoto.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi Drew, I just recently built an Athlon system using the new SOYO Dragon+ MB. So far so good. Works excellent with very good performance. I'm running 2.4.16-xfs with the preempt patches and suffer no ill as a result. I originally started with Mandrake 8.1 then upgraded to 2.4.14 and later to 2.4.16 - It's been working surprisingly well considering the newness of the hardware. System Specs: SOYO Dragon+ MB Athlon XP1800+ 2x256MB Samsung PC2100 DDR 2x40GB IBM GXP60 Deskstar configured with software raid0 GeForce 3 Ti200 Onboard Via ethernet, CM8738 sound both are enabled Additional EEPRO100 All hardware is supported under linux. Using it as my devel workstation/game machine at home and performance is excellent. So far, stability is fine with no lockups, freezes, oopses etc... Been running for approx. 3 weeks. Hope that helps, -Walt Drew 'Cheetah!' Maxwell wrote: > Fellow XFS users: > -------------------------------------------------------------------------------- > My company has been working with XFS-Linux on Intel systems for some > time > and has had some difficulties getting a stable Linux w/XFS release to > work > on some of our older K6-2 systems. Instead of butting our heads against > a > case of potentially suspect hardware, we're looking to upgrade to a > group > of Athlon-based systems. In this, I have a few questions: > > 1. What can those of you running Linux w/XFS recommend for motherboards > to > go with an Athlon processor? > (We might as well get hardware we know works by experience.) > 2. Are they any special changes to your kernels used to work correctly > with > an Athlon-based system? > > Our initial attempts at doing an Athlon + Linux + XFS solution haven't > worked. > We've tried the following methods: > 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using > gcc > 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using > kgcc > Pre-patched XFS 1.0.2 kernel source from SGI, compiled using kgcc > > Any assistance is greatly appreciated. > > --Drew 'Cheetah!' Maxwell [cheetah@clubphoto.com] > ClubPhoto (http://www.clubphoto.com) Pictures, Prints, Gifts, and More! > > From owner-linux-xfs@oss.sgi.com Thu Dec 6 18:47:06 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB72l6h09327 for linux-xfs-outgoing; Thu, 6 Dec 2001 18:47:06 -0800 Received: from starship.berlin (dsl-213-023-043-086.arcor-ip.net [213.23.43.86]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB72l1o09305 for ; Thu, 6 Dec 2001 18:47:03 -0800 Received: from daniel by starship.berlin with local (Exim 3.32 #1 (Debian)) id 16CA5I-0000rw-00; Fri, 07 Dec 2001 02:45:52 +0100 Content-Type: text/plain; charset="iso-8859-1" From: Daniel Phillips To: Nathan Scott , Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher Subject: Re: [PATCH] Revised extended attributes interface Date: Fri, 7 Dec 2001 02:45:50 +0100 X-Mailer: KMail [version 1.3.2] Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <20011207101517.B46546@wobbly.melbourne.sgi.com> In-Reply-To: <20011207101517.B46546@wobbly.melbourne.sgi.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On December 7, 2001 12:15 am, Nathan Scott wrote: > > > (unless you go for a binary namespace representation, and that's a > > > real can of worms). > > > > I'm suggesting we take a look at that. > > Andreas and I did have such an implementation, but we ditched it. > The CVS revision history of cmd/attr2/{set,get}fattr/*.c in the XFS > tree show the progression of user<->kernel interfaces which I tried > while Andreas and I were nutting out a clean solution that we both > could use. > > Thar be dragons thar. Big hairy ones. Could you describe them, please? -- Daniel From owner-linux-xfs@oss.sgi.com Thu Dec 6 19:01:48 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB731m009732 for linux-xfs-outgoing; Thu, 6 Dec 2001 19:01:48 -0800 Received: from starship.berlin (dsl-213-023-043-086.arcor-ip.net [213.23.43.86]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB731io09710 for ; Thu, 6 Dec 2001 19:01:44 -0800 Received: from daniel by starship.berlin with local (Exim 3.32 #1 (Debian)) id 16CAMb-0000s1-00; Fri, 07 Dec 2001 03:03:45 +0100 Content-Type: text/plain; charset="iso-8859-1" From: Daniel Phillips To: Nathan Scott , Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher Subject: Re: [PATCH] Revised extended attributes interface Date: Fri, 7 Dec 2001 03:03:43 +0100 X-Mailer: KMail [version 1.3.2] Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <20011207101517.B46546@wobbly.melbourne.sgi.com> In-Reply-To: <20011207101517.B46546@wobbly.melbourne.sgi.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On December 7, 2001 12:15 am, Nathan Scott wrote: > > > [extended attributes on symlinks] > > > > OK, well it looks like you're going a little overboard here in dividing out > > the functionality. What you're talking about is 'follow symlink or not', > > right? That really does sound to me as though it's naturally expressed with > > a flag bit. I really don't see a compelling reason to go beyond 8 syscalls: > > > > get, fget, set, fset, del, fdel, list, flist > > I'm not too fussed - the second draft patch I sent out did exactly > as you describe, in an attempt to cut down on syscalls. This again > meant adding a "flags" field to each operation. We also have stat/ > lstat/fstat and chown/lchown/fchown - I was trying to be consistent > with those, and I still think that is the right thing to do. It may well be, however, the one call that has flags, set, is looking a little irregular sitting there on its own. We're inventing an API here for which we don't have a lot of guidance from existing unices, correct? It wouldn't hurt to really kick it around. After all, what we settle on in Linux is likely to become the standard. Presumably there's some existing practice at SGI, do you have a pointer to man pages? -- Daniel From owner-linux-xfs@oss.sgi.com Thu Dec 6 19:26:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB73Q3r10143 for linux-xfs-outgoing; Thu, 6 Dec 2001 19:26:03 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB73Puo10119 for ; Thu, 6 Dec 2001 19:25:57 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id SAA03702 for ; Thu, 6 Dec 2001 18:25:48 -0800 (PST) mail_from (lord@sgi.com) Received: from tulip-e185.americas.sgi.com (tulip-e185.americas.sgi.com [128.162.185.208]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id UAA3707983; Thu, 6 Dec 2001 20:24:24 -0600 (CST) Received: from sgi.com (OlmybX1jujWWd9iEE56ZxqGAQSWsIfBr@cf-vpn-sw-corp-64-22.corp.sgi.com [134.15.64.22]) by tulip-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id UAA73747; Thu, 6 Dec 2001 20:24:19 -0600 (CST) Message-ID: <3C1028B6.1010300@sgi.com> Date: Thu, 06 Dec 2001 20:25:58 -0600 From: Stephen Lord User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.6) Gecko/20011120 X-Accept-Language: en-us MIME-Version: 1.0 To: Andi Kleen CC: linux-xfs@oss.sgi.com Subject: Re: 2.5 CVS available References: <1007681057.13482.24.camel@stout.americas.sgi.com> <20011207004053.A22133@wotan.suse.de> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Andi Kleen wrote: >On Thu, Dec 06, 2001 at 05:24:17PM -0600, Eric Sandeen wrote: > >>Yes, that's right, act now to be the first on your block to jeopardize >>your data, and run XFS under Linux 2.5.1-pre5! >> >>Send no money now, simply: >> >>$ export CVSROOT=':pserver:cvs@oss.sgi.com:/cvs' >>$ cvs login (pw: cvs) >>$ cvs -z3 checkout linux-2.5-xfs >> >>Remember, 2.5 is Development with a capital "D" so don't use this for >>anything that _matters_... >> > >Could you give a short summary on the changes? Does it use the new BIO interface yet for >scatter-gather similar to the old kiovec path ? > It uses bio, each pagebuf does its I/O via 1 (or more if too large) bio structures, that is about it for 2.5 specific changes. I have been working with Jens on this as XFS appears to be hitting a couple of bugs in request merging - probably because it is the only serious user of larger bio structs. Because of this there is a good chance of crashes with this tree right now, it is strictly developer only material at the moment. The code base is 2.5.1-pre5 + another bio patch from Jens, the XFS in this tree is actually about a week out of date, I will merge over relevant XFS changes at some point soon. The same commands can be used with the 2.5 kernel, we do not plan on providing two different sets. > > > >-Andi > Steve From owner-linux-xfs@oss.sgi.com Thu Dec 6 20:22:57 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB74Mv611052 for linux-xfs-outgoing; Thu, 6 Dec 2001 20:22:57 -0800 Received: from fnal.gov (heffalump.fnal.gov [131.225.9.20]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB74Mjo11030 for ; Thu, 6 Dec 2001 20:22:45 -0800 Received: from arkroyal-clued0.fnal.gov ([131.225.225.92]) by smtp.fnal.gov (PMDF V6.0-24 #37519) with ESMTP id <0GNY00JNOEPVK8@smtp.fnal.gov> for linux-xfs@oss.sgi.com; Thu, 06 Dec 2001 21:22:43 -0600 (CST) Date: Thu, 06 Dec 2001 21:22:43 -0600 (CST) From: vkuznet Subject: Re: XFS and quota problem In-reply-to: <20011207102411.C46546@wobbly.melbourne.sgi.com> To: Nathan Scott Cc: linux-xfs@oss.sgi.com Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi Nathan, I followed your suggestion and it doesn't work. I download quota-tools from sourceforge, compile them and get the following: vkuznet@vk-clued0(21:01:13)> ./quota -v quota: Error while getting quota from /d02ka.fnal.gov:/desktop/home5/vkuznet for 1212: Connection refused quota: Error while getting quota from /plum-clued0:/disk2/fnal/ups for 1212: Connection refused quota: Error while getting quota from /plum-clued0.fnal.gov:/disk2/local for 1212: Connection refused Disk quotas for user vkuznet (uid 1212): none Any firther suggestions to try? Just to be complete, I use 2.4.14 vanila kernel patched with linux-2.4.14-xfs-2001-11-06.patch and linux-2.4.3-irixnfs-0.1.patch. Valentin. On Fri, 7 Dec 2001, Nathan Scott wrote: > hi Valentin, > > On Thu, Dec 06, 2001 at 11:15:15AM -0600, vkuznet wrote: > > I have a problem with quota command on XFS partitions. > > We have a cluster of ~150 nodes with XFS fs and after > > upgrading to 2.4.14 kernel our quota -v refuse to > > work. The source RPM for quota package from SGI (found on > > rpmfind.net) compiled successfully but show 0 as your > > That is actually a fairly old version of the code now, > you'll see we no longer ship that in recent releases. > You can use the quota code that comes with Redhat now > (since 7.2 IIRC), or download the source from quota-tools > project on sourceforge. > > I have a vague recollection of reports of this problem being > fixed by an upgrade - if not, let me know what happens with > new tools & we'll take it from there. > > cheers. > > > quota limit. Example: > > > > > vkuznet@vk-clued0(12:40:37)> usr/bin/quota -v > > > Disk quotas for user vkuznet (uid 1212): > > > Filesystem blocks quota limit grace files quota limit > > > grace > > > /d02ka.fnal.gov:/desktop/home5/vkuznet > > > 0 0 0 0 0 0 > > > /plum-clued0:/disk2/fnal/ups > > > 0 0 0 0 0 0 > > > /plum-clued0.fnal.gov:/disk2/local > > > 0 0 0 0 0 0 > > > > We do have a variety of disks (local, cross mounted from Linux and > > from SGI file server) and none of them show valid values for quotas. > > > > Can anyone explain the problem and/or provide or point me to valid patch. > > > > I'm not on XFS maling list (sorry too many mails) and will appreciate > > if you replay to my Email directly. > > > > Thanks, > > Valentin. > > > > -- > Nathan > From owner-linux-xfs@oss.sgi.com Thu Dec 6 20:44:06 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB74i6311509 for linux-xfs-outgoing; Thu, 6 Dec 2001 20:44:06 -0800 Received: from rj.sgi.com (rj.sgi.com [204.94.215.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB74i2o11487 for ; Thu, 6 Dec 2001 20:44:02 -0800 Received: from ledzep.americas.sgi.com (ledzep.americas.sgi.com [137.38.226.97]) by rj.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB73huY13316 for ; Thu, 6 Dec 2001 19:43:56 -0800 Received: from maine.americas.sgi.com (maine.americas.sgi.com [128.162.191.42]) by ledzep.americas.sgi.com (SGI-SGI-8.9.3/americas-smart-nospam1.1) with ESMTP id VAA48333 for ; Thu, 6 Dec 2001 21:40:14 -0600 (CST) Received: from nstraz by maine.americas.sgi.com with local (Exim 3.33 #1 (Debian)) id 16CBuJ-0004bT-00 for ; Thu, 06 Dec 2001 21:42:39 -0600 Date: Thu, 6 Dec 2001 21:42:39 -0600 From: Nathan Straz To: linux-xfs@oss.sgi.com Subject: Re: XFS (Linux) and locking problems Message-ID: <20011206214239.A17674@sgi.com> Mail-Followup-To: linux-xfs@oss.sgi.com References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.3.23i Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, Dec 06, 2001 at 03:11:30PM -0800, Dan Rich wrote: > I am seeing odd problems on a Linux XFS filesystem that appear to be > caused by a locking issue. I have had several files dissapear and I > believe it is due to two processes (mail filters) opening the file at the > same time and steping on each other's toes. Have you seen anything like > this on other systems? Do you mean procmail? Are you using a lock file in your recipe? > I'm seeing this with the XFS 1.0.1 release on the 2.4.9 kernel. 2.4.9 was the Redhat based kernel in 1.0.2. 1.0.1 came with a 2.4.3 redhat and a 2.4.5 linus kernel. > Also, is there a timetable for releaseing XFS patches for 2.4.16? They have been released. Check the FTP site. -- Nate Straz nstraz@sgi.com sgi, inc http://www.sgi.com/ Linux Test Project http://ltp.sf.net/ From owner-linux-xfs@oss.sgi.com Thu Dec 6 20:50:45 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB74oj311698 for linux-xfs-outgoing; Thu, 6 Dec 2001 20:50:45 -0800 Received: from zok.sgi.com (zok.SGI.COM [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB74odo11676 for ; Thu, 6 Dec 2001 20:50:39 -0800 Received: from boing.melbourne.sgi.com (boing.melbourne.sgi.com [134.14.55.141]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB73oVA15330 for ; Thu, 6 Dec 2001 19:50:31 -0800 Received: (from tes@localhost) by boing.melbourne.sgi.com (SGI-8.9.3/8.9.3) id OAA52196; Fri, 7 Dec 2001 14:49:06 +1100 (AEDT) Date: Fri, 7 Dec 2001 14:49:05 +1100 From: Timothy Shimmin To: "Quang Nguyen (Ngo)" Cc: "''linux-xfs@oss.sgi.com' '" Subject: Re: attr_set() Changes ctime Message-ID: <20011207144905.M61575@boing.melbourne.sgi.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0us In-Reply-To: ; from quang.nguyen@tapeware.com on Thu, Dec 06, 2001 at 05:18:24PM -0800 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi Quang, On Thu, Dec 06, 2001 at 05:18:24PM -0800, Quang Nguyen (Ngo) wrote: > Does anyone have an answer to this question? ;-) Yep. > > Thanks, > Quang > > -----Original Message----- > From: Quang Nguyen (Ngo) > To: 'linux-xfs@oss.sgi.com' > Sent: 12/6/01 11:21 AM > Subject: attr_set() Changes ctime > > How do I prevent attr_set() from changing the ctime? > I have a feeling that the answer is "you can't". Yep, you can't. The relevant code in xfs_attr_set() is: if (!error && (flags & ATTR_KERNOTIME) == 0) { xfs_ichgtime(dp, XFS_ICHGTIME_CHG); } So one would need ATTR_KERNNOTIME set on the flags for the setting of the attribute. This is done by the dmapi code in the kernel. HOWEVER, it can't be done in userspace as we test to make sure that only valid flags are possibly set: i.e ATTR_ROOT, ATTR_DONTFOLLOW, and for set op: ATTR_CREATE, ATTR_REPLACE linvfs_attrctl(): ... int flags = ops[i].flags; /* common flags */ flags &= ~(ATTR_ROOT | ATTR_DONTFOLLOW); /* command specific */ if (ops[i].opcode == ATTR_OP_SET) flags &= ~(ATTR_CREATE | ATTR_REPLACE); if (flags != 0x0) return -EINVAL; > utime() allows to change mtime and atime of a file. > Is there a function to change the ctime especially for XFS? Not that I know of ?? --Tim From owner-linux-xfs@oss.sgi.com Thu Dec 6 20:53:10 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB74rAV11833 for linux-xfs-outgoing; Thu, 6 Dec 2001 20:53:10 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB74qxo11811 for ; Thu, 6 Dec 2001 20:52:59 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via SMTP id TAA07176 for ; Thu, 6 Dec 2001 19:52:44 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id OAA21725; Fri, 7 Dec 2001 14:51:34 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id OAA54433; Fri, 7 Dec 2001 14:51:32 +1100 (AEDT) Date: Fri, 7 Dec 2001 14:51:31 +1100 From: Nathan Scott To: Daniel Phillips Cc: Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [PATCH] Revised extended attributes interface Message-ID: <20011207145131.B54252@wobbly.melbourne.sgi.com> References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <20011207101517.B46546@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from phillips@bonn-fries.net on Fri, Dec 07, 2001 at 03:03:43AM +0100 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk hi Daniel, On Fri, Dec 07, 2001 at 03:03:43AM +0100, Daniel Phillips wrote: > On December 7, 2001 12:15 am, Nathan Scott wrote: > > > > [extended attributes on symlinks] > > > get, fget, set, fset, del, fdel, list, flist > > > > I'm not too fussed - the second draft patch I sent out did exactly > > as you describe, in an attempt to cut down on syscalls. This again > > meant adding a "flags" field to each operation. We also have stat/ > > lstat/fstat and chown/lchown/fchown - I was trying to be consistent > > with those, and I still think that is the right thing to do. > > It may well be, however, the one call that has flags, set, is looking a > little irregular sitting there on its own. Not sure what to say to that ... the API is practical, flags seem to make sense for that call (the flags give the slightly different "set" semantics, but it is still "set"), IMO they don't make sense for others. > We're inventing an API here for which we don't have a lot of guidance from > existing unices, correct? No. Many existing versions of Unix support extended attributes in some form or another, but there is no common API/standard - each implementation differs, sometimes radically, to the others. > It wouldn't hurt to really kick it around. Please read the archives first - discussion started well over a year ago now on an API for Linux, and there have been heaps and heaps of ideas, proposals, prototypes, etc, floated. The lack of progress on getting something in the kernel is hurting several projects (even having syscalls reserved would be a _huge_ help to us). We have distributors telling us they would include XFS in their kernels if there was some progress on this particular issue. > After all, what we settle on in Linux is likely to become the standard. Mmm.. I seriously doubt there could ever be any standards in this area - I would be satisfied with a good implementation on Linux, which allows filesystems from different operating systems to use it while preserving any existing on-disk formats they may have. > > Presumably there's some existing practice at SGI, do you have a pointer to > man pages? > Start with the mail we sent to Linus, LKML, fs-devel, etc about a month ago - it had pointers to the original discussion from this time last year (and in that discussion, Andreas provided pointers to documentation for several other implementations, incl. BSD, IRIX, Tru64, & others). It also has pointers to several projects relying on the existing, diverging Linux implementations. It should probably be pointed out again that many of the folk working on filesystems which support extended attributes have given their collective thumbs up to the latest round of patches. In particular, the projects which have already implemented EAs (XFS, ext2/ext3) and services above EAs, are confident that these patches will work for them and that we'll be able to get our userspace code working together for the first time. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Thu Dec 6 20:58:39 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB74wdQ12048 for linux-xfs-outgoing; Thu, 6 Dec 2001 20:58:39 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB74wYo12026 for ; Thu, 6 Dec 2001 20:58:34 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via SMTP id TAA06179 for ; Thu, 6 Dec 2001 19:58:35 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id OAA21761; Fri, 7 Dec 2001 14:57:01 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id OAA54520; Fri, 7 Dec 2001 14:57:00 +1100 (AEDT) Date: Fri, 7 Dec 2001 14:56:58 +1100 From: Nathan Scott To: vkuznet Cc: linux-xfs@oss.sgi.com Subject: Re: XFS and quota problem Message-ID: <20011207145658.C54252@wobbly.melbourne.sgi.com> References: <20011207102411.C46546@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from vkuznet@fnal.gov on Thu, Dec 06, 2001 at 09:22:43PM -0600 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, Dec 06, 2001 at 09:22:43PM -0600, vkuznet wrote: > Hi Nathan, hi, > I followed your suggestion and it doesn't work. > I download quota-tools from sourceforge, compile them > and get the following: > > vkuznet@vk-clued0(21:01:13)> ./quota -v > quota: Error while getting quota from > /d02ka.fnal.gov:/desktop/home5/vkuznet for 1212: Connection refused > quota: Error while getting quota from > /plum-clued0:/disk2/fnal/ups for 1212: Connection refused > quota: Error while getting quota from > /plum-clued0.fnal.gov:/disk2/local for 1212: Connection refused > Disk quotas for user vkuznet (uid 1212): none Is rpc.rquotad on these machines running and allowing connections from your machine? (these are NFS mounts right?) "Connection refused" suggests the daemon isn't running on the NFS server. [I have tested this before and it is known to work, so its not that you're doing something new here] > Any firther suggestions to try? > > Just to be complete, I use 2.4.14 vanila kernel > patched with linux-2.4.14-xfs-2001-11-06.patch and > linux-2.4.3-irixnfs-0.1.patch. OK, thanks - that looks like something that should work. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Thu Dec 6 23:37:46 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB77bkt14777 for linux-xfs-outgoing; Thu, 6 Dec 2001 23:37:46 -0800 Received: from dream.lapseofthought.com (adsl-66-124-80-202.dsl.snfc21.pacbell.net [66.124.80.202]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB77beo14755 for ; Thu, 6 Dec 2001 23:37:40 -0800 Received: (from drich@localhost) by dream.lapseofthought.com (8.11.6/8.11.6) id fB76bX603057; Thu, 6 Dec 2001 22:37:33 -0800 Date: Thu, 6 Dec 2001 22:37:33 -0800 (PST) From: Dan Rich X-X-Sender: To: Eric Sandeen cc: Subject: Re: XFS (Linux) and locking problems In-Reply-To: <1007680518.13468.22.camel@stout.americas.sgi.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On 6 Dec 2001, Eric Sandeen wrote: > On Thu, 2001-12-06 at 17:11, Dan Rich wrote: > > I am seeing odd problems on a Linux XFS filesystem that appear to be > > caused by a locking issue. I have had several files dissapear and I > > believe it is due to two processes (mail filters) opening the file at the > > same time and steping on each other's toes. Have you seen anything like > > this on other systems? I'm seeing this with the XFS 1.0.1 release on > > the 2.4.9 kernel. > > Hm, there was no 1.0.1 release for the 2.4.9 kernel... What are the patches in ftp://oss.sgi.com/projects/xfs/download/patches/2.4.9/? > Disappear... as in, no longer in the directory listing? Or contents > disappear? I have seen both happen. The first problem I noticed was a mailbox that ended up with only a few messages in it and all of the old messages missing. I have also seen files that showed in the directory inode but not in the filesystem, i.e. running ls on a directory would return a list of file not found errors followed by a few of the files that were left in the directory. I have also seen pine hang while trying to read a mailbox, I assume becuase it was trying to get a lock on a file (in one case I managed to get it to sieze the lock eventually). > > I haven't tested the 1.0.2 release yet due to problems with the 2.4.14 > > kernel and the 1.0.2 build on my system. > > > > Also, is there a timetable for releaseing XFS patches for 2.4.16? > > They're available in the patches/ dir as of last night or so. The only thing in ftp://oss.sgi.com/projects/xfs/download/Release-1.0.2/kernel_patches/i386/ at the moment is the 2.4.14 patches. Interestingly enough, I now see them in the directory above where I found the 2.4.9 patches. -- Dan Rich | http://www.employees.org/~drich/ | "Step up to red alert!" "Are you sure, sir? | It means changing the bulb in the sign..." | - Red Dwarf (BBC) From owner-linux-xfs@oss.sgi.com Fri Dec 7 03:44:21 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7BiLm27003 for linux-xfs-outgoing; Fri, 7 Dec 2001 03:44:21 -0800 Received: from mail.wzptt.zj.cn (mail.wzptt.zj.cn [202.96.106.130]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7BiGo26979 for ; Fri, 7 Dec 2001 03:44:16 -0800 Date: Fri, 7 Dec 2001 03:44:16 -0800 Message-Id: <200112071144.fB7BiGo26979@oss.sgi.com> Received: from localhost([61.174.184.233]) by mail.wzptt.zj.cn(JetMail 2.5.3.0) with SMTP id jm103c10e960; Fri, 7 Dec 2001 10:39:50 -0000 MIME-Version: 1.0 From: chnze@mail.wzptt.zj.cn Content-Transfer-Encoding: 8bit X-Priority: 1 Subject: Supply electric applance Content-Type: text/plain Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk From: Mr. Sander Yuan Chnze Electric Equipments Co.,Ltd. No.17-505 Lumingyuan Lucheng Industrial Zone Wenzhou China 325007 Fax: 0086-577-88776860 Tel: 0086-577-88776861 or 88776862 E-mail: qunze@mail.wzptt.zj.cn, chnze@mail.wzptt.zj.cn URL: http://www.chnze.com http://www.electricbase.com Dear Sir, We owe your name and address to the internet,and are glad to take this opportunity to introduce our company as a leading manufacturer and exporter for low voltage electrical appliances. We have experiences of exported our products to East-south Asia, Europe, South America, Africa etc of more than five years and received very good and general reputation. In the meantime, our superior products quality and best services gain customers?¡¥s reliability in the world wide. We have over fifty member factories, the products is up to 70 series and more than 8,000 industrial components. Annual production is 20 million sets of the products which have already passed ISO9001 certificates, some products have UL, CSA certificate etc. Our products are in the following: 1. Circuit breaker (MCB,ELCB,MCCB) 2. Ac contactor, Magnetic starter 3. Relays (Mini relay, time relay, thermal relay) 4. Meters(panel meter, water meter, watthour meters) 5. Fuses link and fuse base 6. Stablizer, UPS(uninterruptible power supply) 7. Energy saving lamps 8. Nylon binging tie 9. Micro switch & Limit switch, pushbutton switch 10. Electrical accessories If you want to learn more our products, please kindly let us know, we will send you more information and catalogue for your references. Now we are learned that you are interested in electrical appliances, please send us the details, so that we will send you our favorable prices immediatedly. Best regards Sander Yuan/G.manager http://www.chnze.com --------------------------- ±¾ÓʼþÓÉÍø·Éñ²îÈí¼þ·¢ËÍ. http://member1.shangdu.net/home1/yyang/ From owner-linux-xfs@oss.sgi.com Fri Dec 7 04:35:51 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7CZpL29936 for linux-xfs-outgoing; Fri, 7 Dec 2001 04:35:51 -0800 Received: from mail.ocs.com.au (mail.ocs.com.au [203.34.97.2]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7CZio29914 for ; Fri, 7 Dec 2001 04:35:45 -0800 Received: (qmail 4982 invoked from network); 7 Dec 2001 11:35:40 -0000 Received: from ocs3.intra.ocs.com.au (192.168.255.3) by mail.ocs.com.au with SMTP; 7 Dec 2001 11:35:40 -0000 Received: by ocs3.intra.ocs.com.au (Postfix, from userid 16331) id 73237300090; Fri, 7 Dec 2001 22:35:38 +1100 (EST) Received: from ocs3.intra.ocs.com.au (localhost [127.0.0.1]) by ocs3.intra.ocs.com.au (Postfix) with ESMTP id 28C9696; Fri, 7 Dec 2001 22:35:38 +1100 (EST) X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 From: Keith Owens To: Dan Rich Cc: linux-xfs@oss.sgi.com Subject: Re: XFS (Linux) and locking problems In-reply-to: Your message of "Thu, 06 Dec 2001 22:37:33 -0800." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Fri, 07 Dec 2001 22:35:32 +1100 Message-ID: <27476.1007724932@ocs3.intra.ocs.com.au> Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Thu, 6 Dec 2001 22:37:33 -0800 (PST), Dan Rich wrote: >On 6 Dec 2001, Eric Sandeen wrote: >> Hm, there was no 1.0.1 release for the 2.4.9 kernel... > >What are the patches in ftp://oss.sgi.com/projects/xfs/download/patches/2.4.9/? >From the README (you did read the README, right?): ----- This directory contains a snapshot of the XFS CVS tree at the start of kernel 2.4.9. The snapshot can be downloaded as a single patch containing all the XFS kernel code, it has also been split into several patches to make it easier to port XFS to other architectures and for distributors to include XFS in their distributions. You can apply the -all patch to get a snapshot of XFS for 2.4.9-i386 as of 2001-08-22 04:58 UTC. Or you can selectively apply the -split patches. Do not mix the -all and -split patches, that will result in patch errors and/or duplicate code. The -all and -split patches are provided as a courtesy and are not supported. They are created after the initial upgrade to a new kernel and are not changed afterwards. In particular any bug fixes or XFS/kdb changes for pre-release kernels are _not_ reflected here. The main method of XFS distribution is via the XFS CVS tree, if you want an up to date version of XFS then use CVS. ----- Snapshot != official XFS release. From owner-linux-xfs@oss.sgi.com Fri Dec 7 11:48:14 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7JmEK20129 for linux-xfs-outgoing; Fri, 7 Dec 2001 11:48:14 -0800 Received: from clubphoto.com (mail-gw.clubphoto.com [64.124.34.66]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7Jm5o20106 for ; Fri, 7 Dec 2001 11:48:06 -0800 Received: from [192.168.168.125] ([192.168.168.125]) by clubphoto.com (8.9.3/8.9.3) with ESMTP id KAA25942; Fri, 7 Dec 2001 10:47:58 -0800 User-Agent: Microsoft-Entourage/10.0.0.1309 Date: Fri, 07 Dec 2001 10:48:10 -0800 Subject: Re: Linux w/XFS on Athlon systems From: "Gabe E. Nydick" To: Walt H , "Drew 'Cheetah!' Maxwell" CC: Message-ID: In-Reply-To: <3C101BAB.1060901@mindspring.com> Mime-version: 1.0 Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk I looked yesterday for the 2.4.16 patches, couldn't find them, and I couldn't find the preempt patches either, can you point me to them? Thanks, --------------------- Gabe E. Nydick Systems Lead ClubPhoto, Inc. P: 408.423.6611 F: 408.557.6799 ---------------------- On 12/6/01 5:30 PM, "Walt H" wrote: > Hi Drew, > > I just recently built an Athlon system using the new SOYO Dragon+ MB. So > far so good. Works excellent with very good performance. I'm running > 2.4.16-xfs with the preempt patches and suffer no ill as a result. I > originally started with Mandrake 8.1 then upgraded to 2.4.14 and later > to 2.4.16 - It's been working surprisingly well considering the newness > of the hardware. > > System Specs: > SOYO Dragon+ MB > Athlon XP1800+ > 2x256MB Samsung PC2100 DDR > 2x40GB IBM GXP60 Deskstar configured with software raid0 > GeForce 3 Ti200 > Onboard Via ethernet, CM8738 sound both are enabled > Additional EEPRO100 > All hardware is supported under linux. > > Using it as my devel workstation/game machine at home and performance is > excellent. So far, stability is fine with no lockups, freezes, oopses > etc... Been running for approx. 3 weeks. Hope that helps, > > -Walt > > > Drew 'Cheetah!' Maxwell wrote: > >> Fellow XFS users: >> ----------------------------------------------------------------------------- >> --- >> My company has been working with XFS-Linux on Intel systems for some >> time >> and has had some difficulties getting a stable Linux w/XFS release to >> work >> on some of our older K6-2 systems. Instead of butting our heads against >> a >> case of potentially suspect hardware, we're looking to upgrade to a >> group >> of Athlon-based systems. In this, I have a few questions: >> >> 1. What can those of you running Linux w/XFS recommend for motherboards >> to >> go with an Athlon processor? >> (We might as well get hardware we know works by experience.) >> 2. Are they any special changes to your kernels used to work correctly >> with >> an Athlon-based system? >> >> Our initial attempts at doing an Athlon + Linux + XFS solution haven't >> worked. >> We've tried the following methods: >> 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using >> gcc >> 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using >> kgcc >> Pre-patched XFS 1.0.2 kernel source from SGI, compiled using kgcc >> >> Any assistance is greatly appreciated. >> >> --Drew 'Cheetah!' Maxwell [cheetah@clubphoto.com] >> ClubPhoto (http://www.clubphoto.com) Pictures, Prints, Gifts, and More! >> >> > From owner-linux-xfs@oss.sgi.com Fri Dec 7 11:58:12 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7JwCA20518 for linux-xfs-outgoing; Fri, 7 Dec 2001 11:58:12 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7Jw6o20486 for ; Fri, 7 Dec 2001 11:58:06 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id TAA1674234 for ; Fri, 7 Dec 2001 19:58:00 +0100 (CET) mail_from (sandeen@sgi.com) Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id MAA3733088; Fri, 7 Dec 2001 12:56:39 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id MAA35887; Fri, 7 Dec 2001 12:56:38 -0600 (CST) Subject: Re: Linux w/XFS on Athlon systems From: Eric Sandeen To: "Gabe E. Nydick" Cc: Walt H , "Drew 'Cheetah!' Maxwell" , linux-xfs@oss.sgi.com In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 07 Dec 2001 12:56:29 -0600 Message-Id: <1007751390.26739.10.camel@stout.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Fri, 2001-12-07 at 12:48, Gabe E. Nydick wrote: > I looked yesterday for the 2.4.16 patches, couldn't find them, and I > couldn't find the preempt patches either, can you point me to them? ftp://oss.sgi.com/projects/xfs/download/patches/2.4.16/ preempt patches: http://www.tech9.net/rml/linux/ -Eric -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Fri Dec 7 12:05:56 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7K5uL20812 for linux-xfs-outgoing; Fri, 7 Dec 2001 12:05:56 -0800 Received: from clubphoto.com (mail-gw.clubphoto.com [64.124.34.66]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7K5qo20789 for ; Fri, 7 Dec 2001 12:05:52 -0800 Received: from [192.168.168.125] ([192.168.168.125]) by clubphoto.com (8.9.3/8.9.3) with ESMTP id LAA26209; Fri, 7 Dec 2001 11:05:49 -0800 User-Agent: Microsoft-Entourage/10.0.0.1309 Date: Fri, 07 Dec 2001 11:06:01 -0800 Subject: Re: Linux w/XFS on Athlon systems From: "Gabe E. Nydick" To: Eric Sandeen CC: Walt H , "Drew 'Cheetah!' Maxwell" , Message-ID: In-Reply-To: <1007751390.26739.10.camel@stout.americas.sgi.com> Mime-version: 1.0 Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Eric, Is the preempt patch supposed to fix anything with athlons? Thanks, --------------------- Gabe E. Nydick Systems Lead ClubPhoto, Inc. P: 408.423.6611 F: 408.557.6799 ---------------------- On 12/7/01 10:56 AM, "Eric Sandeen" wrote: > On Fri, 2001-12-07 at 12:48, Gabe E. Nydick wrote: >> I looked yesterday for the 2.4.16 patches, couldn't find them, and I >> couldn't find the preempt patches either, can you point me to them? > > ftp://oss.sgi.com/projects/xfs/download/patches/2.4.16/ > > preempt patches: > http://www.tech9.net/rml/linux/ > > -Eric From owner-linux-xfs@oss.sgi.com Fri Dec 7 12:15:48 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7KFmv21243 for linux-xfs-outgoing; Fri, 7 Dec 2001 12:15:48 -0800 Received: from zok.sgi.com (zok.sgi.com [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7KFjo21221 for ; Fri, 7 Dec 2001 12:15:45 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB7JFdA29805 for ; Fri, 7 Dec 2001 11:15:39 -0800 Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id NAA3782956 for ; Fri, 7 Dec 2001 13:14:23 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id NAA80987 for ; Fri, 7 Dec 2001 13:14:23 -0600 (CST) From: Steve Lord Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB7JDgj32496; Fri, 7 Dec 2001 13:13:42 -0600 Message-Id: <200112071913.fB7JDgj32496@jen.americas.sgi.com> Date: Fri, 7 Dec 2001 13:13:42 -0600 Subject: TAKE - add new kdb command Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Date: Fri Dec 7 11:14:09 PST 2001 Workarea: jen.americas.sgi.com:/src/lord/xfs-linux.2.4 The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs Modid: 2.4.x-xfs:slinx:108022a linux/kdb/modules/kdbm_pg.c - 1.45 - Add a new command for dumping page state summary From owner-linux-xfs@oss.sgi.com Fri Dec 7 12:20:31 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7KKVV21487 for linux-xfs-outgoing; Fri, 7 Dec 2001 12:20:31 -0800 Received: from zok.sgi.com (zok.sgi.com [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7KKPo21465 for ; Fri, 7 Dec 2001 12:20:25 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB7JKJA30208 for ; Fri, 7 Dec 2001 11:20:19 -0800 Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id NAA3795203 for ; Fri, 7 Dec 2001 13:19:03 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id NAA81140 for ; Fri, 7 Dec 2001 13:19:03 -0600 (CST) From: Steve Lord Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB7JIM632566; Fri, 7 Dec 2001 13:18:22 -0600 Message-Id: <200112071918.fB7JIM632566@jen.americas.sgi.com> Date: Fri, 7 Dec 2001 13:18:22 -0600 Subject: TAKE - merge up to 2.5.1-pre6 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk I cannot find anything significant in the changes, but I cannot break this version. This of course means that the only way to break it is to push it out onto oss ;-) Steve Date: Fri Dec 7 11:17:45 PST 2001 Workarea: jen.americas.sgi.com:/src/lord/xfs-merge The following file(s) were checked into: bonnie.engr.sgi.com:/isms/slinx/2.5.x-xfs Modid: 2.5.x-xfs:slinx:108023a linux/include/linux/coda_linux.h - 1.12 linux/include/linux/blkdev.h - 1.45 linux/include/linux/blk.h - 1.23 linux/include/asm-alpha/page.h - 1.13 linux/fs/coda/upcall.c - 1.14 linux/fs/coda/sysctl.c - 1.13 linux/fs/coda/psdev.c - 1.17 linux/fs/coda/pioctl.c - 1.10 linux/fs/coda/inode.c - 1.14 linux/fs/coda/file.c - 1.16 linux/fs/coda/dir.c - 1.20 linux/fs/coda/cnode.c - 1.12 linux/fs/coda/cache.c - 1.9 linux/drivers/scsi/sr.c - 1.32 linux/drivers/isdn/avmb1/capi.c - 1.26 linux/drivers/cdrom/cdrom.c - 1.38 linux/drivers/block/nbd.c - 1.26 linux/drivers/block/ll_rw_blk.c - 1.82 linux/drivers/block/genhd.c - 1.22 linux/Makefile - 1.161 linux/MAINTAINERS - 1.86 linux/CREDITS - 1.68 linux/drivers/isdn/eicon/eicon_mod.c - 1.15 linux/drivers/isdn/divert/divert_procfs.c - 1.17 linux/drivers/isdn/hysdn/hysdn_procfs.c - 1.7 linux/drivers/ide/ide-cd.c - 1.21 linux/drivers/media/video/stradis.c - 1.8 linux/drivers/media/video/saa7146.h - 1.2 linux/drivers/isdn/eicon/common.c - 1.8 linux/arch/i386/kernel/dmi_scan.c - 1.12 linux/drivers/isdn/hisax/st5481_b.c - 1.3 linux/drivers/isdn/hisax/st5481_d.c - 1.4 linux/drivers/isdn/hisax/st5481_usb.c - 1.4 linux/include/linux/bio.h - 1.2 linux/drivers/isdn/hisax/hisax_fcpcipnp.c - 1.2 linux/fs/bio.c - 1.2 linux/drivers/block/block_ioctl.c - 1.2 linux/drivers/isdn/hisax/hisax_isac.h - 1.2 linux/drivers/isdn/hisax/hisax_isac.c - 1.2 From owner-linux-xfs@oss.sgi.com Fri Dec 7 12:40:31 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7KeVD22310 for linux-xfs-outgoing; Fri, 7 Dec 2001 12:40:31 -0800 Received: from minnie.omroep.nl (minnie.omroep.nl [145.58.30.4]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7KeQo22285 for ; Fri, 7 Dec 2001 12:40:26 -0800 Received: (from smap@localhost) by minnie.omroep.nl (SGI-8.9.3/8.9.3) id UAA85819 for ; Fri, 7 Dec 2001 20:40:19 +0100 (CET) Received: from kwek.omroep.nl (145.58.31.3 "EHLO kwek.omroep.nl") by minnie.omroep.nl with ESMTP (smap v3.0 nederlandse publieke omroep) id xma5337574; Fri, 7 Dec 01 20:40:15 +0100 Received: from localhost (matthijs@localhost) by kwek.omroep.nl (SGI-8.9.3/8.9.3) with ESMTP id UAA62414 for ; Fri, 7 Dec 2001 20:40:13 +0100 (CET) X-Authentication-Warning: kwek.omroep.nl: matthijs owned process doing -bs Date: Fri, 7 Dec 2001 20:40:13 +0100 From: Matthijs van der Klip X-X-Sender: matthijs@kwek.omroep.nl To: Linux XFS Mailing List Subject: using dmapi to sync filesystems (was using xfsdump...) Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, Some days ago I submitted a crazy idea to sync some filesystems through the use of xfsdump. As I already expected this was not really a viable solution. Someone (Seth Mos) however pointed me to DMAPI. I have already done some research on it but need some info from people who know DMAPI and I think I can find those people on this list... :) I would like to try to create an app that does the following: 1) Register with dmapi. 2) Tell dmapi I want to receive events when directory entries on a certain filesystem are added/deleted. 3) On receipt of an event append the absolute path of the modified directory to a logfile. >From what I've read about DMAPI I think this should be possible and not particularly very hard to achieve even. However I'm not an experienced C programmer, so I'd be thankful if someone would give me some hints, tips, pseudo code or whatever else that can help me. The XDSM standard at http://www.opengroup.org/onlinepubs/9657099/toc.htm tells me to call 'dm_init_service' before doing anything else. I can not find any reference to this function in the source of xfsdump however. Can anyone perhaps give me a short list of needed functions calls needed for above functionality? Thanks in advance, -- Matthijs van der Klip - Unix Administrator NOS - Dutch Public Broadcasting Organisation http://www.omroep.nl From owner-linux-xfs@oss.sgi.com Fri Dec 7 13:11:23 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7LBNs23308 for linux-xfs-outgoing; Fri, 7 Dec 2001 13:11:23 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7LBHo23286 for ; Fri, 7 Dec 2001 13:11:17 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id MAA04533 for ; Fri, 7 Dec 2001 12:11:07 -0800 (PST) mail_from (roehrich@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id OAA3725588; Fri, 7 Dec 2001 14:09:49 -0600 (CST) Received: from slobber.americas.sgi.com (slobber.americas.sgi.com [128.162.187.52]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id OAA84657; Fri, 7 Dec 2001 14:09:49 -0600 (CST) Received: from slobber.americas.sgi.com by slobber.americas.sgi.com (SGI-8.9.3/SGI-client-1.7) via ESMTP id OAA01605; Fri, 7 Dec 2001 14:09:48 -0600 (CST) Message-Id: <200112072009.OAA01605@slobber.americas.sgi.com> To: Matthijs van der Klip cc: Linux XFS Mailing List Subject: Re: using dmapi to sync filesystems (was using xfsdump...) Date: Fri, 07 Dec 2001 14:09:47 -0600 From: Dean Roehrich Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >From: Matthijs van der Klip >I would like to try to create an app that does the following: > >1) Register with dmapi. >2) Tell dmapi I want to receive events when directory entries on a certain > filesystem are added/deleted. >3) On receipt of an event append the absolute path of the modified > directory to a logfile. Using XFS and the dmapi tests in cmd/xfstests/dmapi: # mount -o dmapi /dev/hda12 /mnts/dmi1 # dm_create_session -i dean1 ret=0 newsid=1 # set_disp -s 1 /mnts/dmi1 DM_EVENT_CREATE DM_EVENT_REMOVE # set_eventlist -s 1 /mnts/dmi1 DM_EVENT_CREATE DM_EVENT_REMOVE # touch /mnts/dmi1/foo1& [1] 771 # # get_events 1 rlenp is 96 create: token=1 sequence=1 parent dir 5a659482b0bbecee0e000000000000004000000000000000 name foo1 mode bits mode 100644: perm rw- r-- r--, type Regular File # respond_event 1 1 1 0 # [1] + Done touch /mnts/dmi1/foo1& To get actual pathnames, well that's different. Our HSM doesn't need the pathname, and so...it's not till now that I see maybe dm_handle_to_path() in libdm is broken. I'm sure someone who needs it could come up with a fix easily enough :) Dean From owner-linux-xfs@oss.sgi.com Fri Dec 7 13:18:47 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7LIlr23620 for linux-xfs-outgoing; Fri, 7 Dec 2001 13:18:47 -0800 Received: from mailbeast.ahaza.com (mailbeast.ahaza.com [209.180.221.132]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7LIeo23598 for ; Fri, 7 Dec 2001 13:18:40 -0800 Received: from akira.ahaza.com (akira.ahaza.com [172.16.30.230]) by mailbeast.ahaza.com (8.11.1/8.11.1) with ESMTP id fB7KIWX90247; Fri, 7 Dec 2001 12:18:32 -0800 (PST) (envelope-from harlequin@tgi.net) Received: from tgi.net (localhost.ahaza.com [127.0.0.1]) by akira.ahaza.com (8.11.6/8.11.6) with ESMTP id fB7KJEe12679; Fri, 7 Dec 2001 12:19:15 -0800 (PST) (envelope-from harlequin@tgi.net) Message-ID: <3C112442.3F157596@tgi.net> Date: Fri, 07 Dec 2001 12:19:14 -0800 From: Tim Wiess X-Mailer: Mozilla 4.78 [en] (X11; U; Linux 2.2.12 i386) X-Accept-Language: en MIME-Version: 1.0 To: "Gabe E. Nydick" CC: Seth Mos , "Drew 'Cheetah!' Maxwell" , Jesse Hall , linux-xfs@oss.sgi.com Subject: Re: Linux w/XFS on Athlon systems References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk I recently put together a dual Athlon system, and initially I couldn't even get it to run for more than 30 minutes with a panic. The problem turned out to a lack of power. I replaced the 300W power supply I had with a 400W and now the machine runs beautifully. If you are testing a dual Athlon configuration, I recommend checking that. tim "Gabe E. Nydick" wrote: > > Seth, > > The system we have works fine with 1000 simultaneous bonnie++ processes > running under a PIII kernel. Do you think it could still be the memory? > > Thanks, > Gabe > > On 12/6/01 11:35 AM, "Seth Mos" wrote: > > > On Thu, 6 Dec 2001, Drew 'Cheetah!' Maxwell wrote: > > > >> Jesse Hall wrote: > >> ----------------------------------------------------------------------------- > >> --- > >>> > >>> What do you mean by "haven't worked"? > >>> > >>> I've been using a Tyan Thunder K7 with dual Athlon MPs with XFS (tracking > >>> CVS) for a few months with great reliability and stability. > >>> > >> By "haven't worked", I mean generating a Kernel Oops, and never being able to > >> finish bringing the system up to multiuser mode. (It'll die somewhere in the > >> midst of the interactive menus as it brings up daemons.) > > > > Sounds to me like RAM with a lot of holes in it. > > > > Get some/new other ones and make sure that the processor is not > > overheating and such. > > > > Cheers > > From owner-linux-xfs@oss.sgi.com Fri Dec 7 13:21:24 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7LLO123791 for linux-xfs-outgoing; Fri, 7 Dec 2001 13:21:24 -0800 Received: from sisko.scot.redhat.com (pc-62-31-66-178-ed.blueyonder.co.uk [62.31.66.178]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7LLKo23768 for ; Fri, 7 Dec 2001 13:21:21 -0800 Received: (from sct@localhost) by sisko.scot.redhat.com (8.11.6/8.11.2) id fB7KKaQ05617; Fri, 7 Dec 2001 20:20:36 GMT Date: Fri, 7 Dec 2001 20:20:36 +0000 From: "Stephen C. Tweedie" To: Nathan Scott Cc: Linus Torvalds , Alexander Viro , Andi Kleen , Andreas Gruenbacher , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com, Stephen Tweedie Subject: Re: [PATCH] Revised extended attributes interface Message-ID: <20011207202036.J2274@redhat.com> References: <20011205143209.C44610@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20011205143209.C44610@wobbly.melbourne.sgi.com>; from nathans@sgi.com on Wed, Dec 05, 2001 at 02:32:10PM +1100 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, On Wed, Dec 05, 2001 at 02:32:10PM +1100, Nathan Scott wrote: > Here is the revised interface. I believe it takes into account > the issues raised so far - further suggestions are also welcome, > of course. This is looking OK as far as EAs go. However, there is still no mention of ACLs specifically, except an oblique reference to ""system.posix_acl_access". Is there no consensus on this? In previous proposals we've at least tried to deal with it to some extent. Cheers, Stephen From owner-linux-xfs@oss.sgi.com Fri Dec 7 14:07:50 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7M7ow25033 for linux-xfs-outgoing; Fri, 7 Dec 2001 14:07:50 -0800 Received: from chef.cc.absoval.com (cpe-66-1-218-101.fl.sprintbbd.net [66.1.218.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7M7do25010 for ; Fri, 7 Dec 2001 14:07:39 -0800 Received: from ieee.org (bs@thebs.cc.absoval.com [192.168.100.89]) by chef.cc.absoval.com (8.9.3/8.9.3) with ESMTP id QAA31459; Fri, 7 Dec 2001 16:06:29 -0500 Message-ID: <3C112F72.A6D83320@ieee.org> Date: Fri, 07 Dec 2001 16:06:58 -0500 From: Bryan-TheBS-Smith Organization: SmithConcepts/AbsoluteValueSystems X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.2.20-udma100-ext3 i686) X-Accept-Language: en MIME-Version: 1.0 To: Tim Wiess CC: "Gabe E. Nydick" , Seth Mos , "Drew 'Cheetah!' Maxwell" , Jesse Hall , linux-xfs@oss.sgi.com Subject: Re: Linux w/XFS on Athlon systems References: <3C112442.3F157596@tgi.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Tim Wiess wrote: > I recently put together a dual Athlon system, and initially > I couldn't even get it to run for more than 30 minutes with > a panic. The problem turned out to a lack of power. I replaced > the 300W power supply I had with a 400W and now the machine > runs beautifully. If you are testing a dual Athlon configuration, > I recommend checking that. Yep. I have a dual-Athlon and even though the Antec 412X 400W's specs were well beyond what Tyan said was required for the Tiger MP (the stardard ATX form-factor/PS board), going to an Enermax 465VE 431W with just a few more amps on the 3/5/12V lines made all the difference. Antec is going to release a 510W soon for the dual-Athlon is that tells you anything. ;-PPP Airflow and cooling are other considerations as well. In addition to some extra 80mm in/outtake fans (Antec SX600, 1000, 1200 series are great), the Thermaltake Volcano 6Cu is an oversized sink with a copperbase, and a tall 25mm high (60x60mm) fan that moves 32CFM and is fairly silent (only 4,550rpms, 30db). No need to get the 6Cu+ with the 38CFM (7,000rpm, 39db!). -- TheBS -- Bryan "TheBS" Smith mailto:b.j.smith@ieee.org chat:thebs413 Engineer AbsoluteValue Systems, Inc. http://www.linux-wlan.org President SmithConcepts, Inc. http://www.SmithConcepts.com ---------------------------------------------------------------- "The [US] Constitution guarantees you Free, not Fair. 'Fair' is a socialist concept." -- Shawn McMahon From owner-linux-xfs@oss.sgi.com Fri Dec 7 14:18:26 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7MIQO25333 for linux-xfs-outgoing; Fri, 7 Dec 2001 14:18:26 -0800 Received: from microsharp.com (mail.microsharp.com [206.190.130.135]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7MIKo25311 for ; Fri, 7 Dec 2001 14:18:20 -0800 Received: from juniper.intra.microsharp.com (gw.intra.microsharp.com [206.190.130.132] (may be forged)) by microsharp.com (8.11.2/8.11.2) with ESMTP id fB7LD2H07640 for ; Fri, 7 Dec 2001 13:13:03 -0800 Received: from juniper.intra.microsharp.com (karlheg@localhost [127.0.0.1]) by juniper.intra.microsharp.com (8.12.1/8.12.1/Debian -2) with ESMTP id fB7LI9sX007448 for ; Fri, 7 Dec 2001 13:18:09 -0800 Received: (from karlheg@localhost) by juniper.intra.microsharp.com (8.12.1/8.12.1/Debian -2) id fB7LI9Nw007444; Fri, 7 Dec 2001 13:18:09 -0800 To: linux-xfs@oss.sgi.com Subject: [XFS-1.0.2,LVM-1.0.1,linux-2.4.14] "mount -U " does not work with XFS. X-Face: 1v=v?=]<7%cW&OL:Z"GCcdmIN&wp)w|1EKzyC6?qaX;55wwjPfKTm\'7x-gd+KEd`(o&(EQ%(B4]6sZ$|e_FBGTJOvJtHUM>flSNQqfdw.jYZ$jnCERoAX*{4lEJm?EB#V_<2\-#5%#Du/ExQR)juedX>!Cq#;UoAo?|Oo~BX@!fwIX From: karlheg@microsharp.com (Karl M. Hegbloom) Date: 07 Dec 2001 13:18:08 -0800 Message-ID: <871yi6zq3z.fsf@juniper.intra.microsharp.com> Lines: 25 User-Agent: Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.4 (Artificial Intelligence) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk [ I am not subscribed to the list; Please Cc. ] I'm using "mount" version 2.11m-1 (debian "sid"), Linux 2.4.14 from Debian "kernel-source", version 2.4.14-1, patched with Debian supplied "kernel-patch-xfs", version 1.0.2-2, and LVM release 1.0.1 patched in by myself from source obtained directly from the LVM site... I've set up a VG with one PV in it, and in there have created one LV, and have created an XFS filesystem on that. It shows up in "/proc/partitions". I cannot get "mount -U /mnt/tmp" to work, nor can I get the UUID= style fstab entry to work, for XFS. I created a second LV and made an ext2fs on it, and tried to do a "mount -U" style mount, and _it_ DID work. The XFS there mounts just fine when I specify the device name in the mount command. -- mailto: (Karl M. Hegbloom) karlheg@microsharp.com http://www.microsharp.com phone://USA/WA/360-260-2066 jabber: karlheg@jabber.org From owner-linux-xfs@oss.sgi.com Fri Dec 7 14:20:27 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7MKRf25497 for linux-xfs-outgoing; Fri, 7 Dec 2001 14:20:27 -0800 Received: from stine.vestdata.no (IDENT:0@stine.vestdata.no [195.204.68.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7MKLo25472 for ; Fri, 7 Dec 2001 14:20:21 -0800 Received: (from ragnark@localhost) by stine.vestdata.no (8.11.6/8.11.2) id fB7LJtj18839; Fri, 7 Dec 2001 22:19:55 +0100 Date: Fri, 7 Dec 2001 22:19:55 +0100 From: =?iso-8859-1?Q?Ragnar_Kj=F8rstad?= To: Matthijs van der Klip Cc: Linux XFS Mailing List , braam@clusterfilesystem.com Subject: Re: using dmapi to sync filesystems (was using xfsdump...) Message-ID: <20011207221955.A18104@vestdata.no> References: Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from matthijs.van.der.klip@nos.nl on Fri, Dec 07, 2001 at 08:40:13PM +0100 X-MIME-Autoconverted: from 8bit to quoted-printable by stine.vestdata.no id fB7LJtj18839 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id fB7MKMo25475 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk CC'ed to Peter Braam for inter-mezzo comments. On Fri, Dec 07, 2001 at 08:40:13PM +0100, Matthijs van der Klip wrote: > Some days ago I submitted a crazy idea to sync some filesystems through > the use of xfsdump. As I already expected this was not really a viable > solution. Someone (Seth Mos) however pointed me to DMAPI. I have already > done some research on it but need some info from people who know DMAPI and > I think I can find those people on this list... :) This sounds like a very interesting idea indeed. > 1) Register with dmapi. > 2) Tell dmapi I want to receive events when directory entries on a certain > filesystem are added/deleted. > 3) On receipt of an event append the absolute path of the modified > directory to a logfile. Well, you can find _one_ filename for the object that change - not neccesarily the one the user used. For directories this would be the same, but for files with hardlinks it may be different. You could pretty easily extend your idea to also include getting notified when files are written to, so the data can be synced to your remote server. It is even possible to get a syncronious event before data is read, so your userspace program get a chance to retrieve the data before the read() operation completes. To me this sounds like a really clean way to integrate something like inter-mezzo with a standard filesystem! -- Ragnar Kjørstad Big Storage From owner-linux-xfs@oss.sgi.com Fri Dec 7 15:13:49 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7NDne26725 for linux-xfs-outgoing; Fri, 7 Dec 2001 15:13:49 -0800 Received: from chimta01.algx.net (chimta01.algx.net [216.99.233.34]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7NDko26702 for ; Fri, 7 Dec 2001 15:13:46 -0800 Received: from jtsdell (66-2-81-28.customer.algx.net [66.2.81.28]) by chimmx01.algx.net (iPlanet Messaging Server 5.1 (built May 7 2001)) with ESMTP id <0GNZ00H4QUYN6D@chimmx01.algx.net> for linux-xfs@oss.sgi.com; Fri, 07 Dec 2001 16:11:11 -0600 (CST) Date: Fri, 07 Dec 2001 17:09:07 -0500 (EST) From: jtrostel@snapserver.com Subject: Oops in 2.5.1-pre6 To: XFS list Reply-to: jtrostel@snapserver.com Message-id: Organization: Quantum Corp. / NASD MIME-version: 1.0 X-Mailer: XFMail 1.5.1 on Linux Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7BIT X-Priority: 3 (Normal) Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hmmmm.... don't think this should happen depmod: *** Unresolved symbols in /lib/modules/2.5.1-pre6-xfs/kernel/fs/pagebuf/pagebuf.o depmod: submit_bio -- John M. Trostel Senior Software Engineer Quantum Corp. / NASD jtrostel@snapserver.com From owner-linux-xfs@oss.sgi.com Fri Dec 7 15:30:17 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7NUHs27261 for linux-xfs-outgoing; Fri, 7 Dec 2001 15:30:17 -0800 Received: from zok.sgi.com (zok.sgi.com [204.94.215.101]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7NUDo27239 for ; Fri, 7 Dec 2001 15:30:13 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by zok.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fB7MU7A10310 for ; Fri, 7 Dec 2001 14:30:07 -0800 Received: from poppy-e185.americas.sgi.com (poppy-e185.americas.sgi.com [128.162.185.207]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id QAA3775076; Fri, 7 Dec 2001 16:28:51 -0600 (CST) Received: from stout.americas.sgi.com (stout.americas.sgi.com [128.162.187.5]) by poppy-e185.americas.sgi.com (980427.SGI.8.8.8/SGI-server-1.7) with ESMTP id QAA44820; Fri, 7 Dec 2001 16:28:49 -0600 (CST) Subject: Re: Oops in 2.5.1-pre6 From: Eric Sandeen To: jtrostel@snapserver.com Cc: XFS list In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.99.2 (Preview Release) Date: 07 Dec 2001 16:28:39 -0600 Message-Id: <1007764120.26739.19.camel@stout.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Just toss EXPORT_SYMBOL(submit_bio); in at the end of fs/bio.c. -Eric On Fri, 2001-12-07 at 16:09, jtrostel@snapserver.com wrote: > Hmmmm.... don't think this should happen > > depmod: *** Unresolved symbols in > /lib/modules/2.5.1-pre6-xfs/kernel/fs/pagebuf/pagebuf.o > depmod: submit_bio > > -- > John M. Trostel > Senior Software Engineer > Quantum Corp. / NASD > jtrostel@snapserver.com -- Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs sandeen@sgi.com SGI, Inc. From owner-linux-xfs@oss.sgi.com Fri Dec 7 15:35:48 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB7NZmA27528 for linux-xfs-outgoing; Fri, 7 Dec 2001 15:35:48 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB7NZho27503 for ; Fri, 7 Dec 2001 15:35:43 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id OAA02259 for ; Fri, 7 Dec 2001 14:35:40 -0800 (PST) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id QAA3802880; Fri, 7 Dec 2001 16:34:24 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id QAA92918; Fri, 7 Dec 2001 16:34:24 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fB7MXff05949; Fri, 7 Dec 2001 16:33:41 -0600 Subject: Re: Oops in 2.5.1-pre6 From: Steve Lord To: Eric Sandeen Cc: jtrostel@snapserver.com, XFS list In-Reply-To: <1007764120.26739.19.camel@stout.americas.sgi.com> References: <1007764120.26739.19.camel@stout.americas.sgi.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 07 Dec 2001 16:33:41 -0600 Message-Id: <1007764421.32585.8.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Fri, 2001-12-07 at 16:28, Eric Sandeen wrote: > Just toss > > EXPORT_SYMBOL(submit_bio); > > in at the end of fs/bio.c. Actually drivers/block/ll_rw_blk.c - or build pagebuf into the kernel. Steve > > -Eric > > On Fri, 2001-12-07 at 16:09, jtrostel@snapserver.com wrote: > > Hmmmm.... don't think this should happen > > > > depmod: *** Unresolved symbols in > > /lib/modules/2.5.1-pre6-xfs/kernel/fs/pagebuf/pagebuf.o > > depmod: submit_bio > > > > -- > > John M. Trostel > > Senior Software Engineer > > Quantum Corp. / NASD > > jtrostel@snapserver.com > -- > Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs > sandeen@sgi.com SGI, Inc. -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Fri Dec 7 19:16:25 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB83GPo31642 for linux-xfs-outgoing; Fri, 7 Dec 2001 19:16:25 -0800 Received: from blount.mail.mindspring.net (blount.mail.mindspring.net [207.69.200.226]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB83GJo31620 for ; Fri, 7 Dec 2001 19:16:19 -0800 Received: from user-uini69j.dsl.mindspring.com ([165.121.25.51] helo=waltsathlon.localhost.net) by blount.mail.mindspring.net with esmtp (Exim 3.33 #1) id 16CX2G-000225-00 for linux-xfs@oss.sgi.com; Fri, 07 Dec 2001 21:16:16 -0500 Received: from mindspring.com (localhost.localdomain [127.0.0.1]) by waltsathlon.localhost.net (Postfix) with ESMTP id ED648D03A217; Fri, 7 Dec 2001 18:15:29 -0800 (PST) Message-ID: <3C1177C1.2070004@mindspring.com> Date: Fri, 07 Dec 2001 18:15:29 -0800 From: Walt H User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.6) Gecko/20011120 X-Accept-Language: en-us MIME-Version: 1.0 To: "Gabe E. Nydick" Cc: Eric Sandeen , "Drew 'Cheetah!' Maxwell" , linux-xfs@oss.sgi.com Subject: Re: Linux w/XFS on Athlon systems References: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Actually, they aren't architecture specific, they are "just" a set of patches which allow the linux kernel to be preempted for higher priority tasks (I hope I didn't mangle that definition too much). At any rate, I find that doing things such as long compiles, while simultaneously playing mp3's etc... seem to work better with them installed. It makes my computer feel snappier overall :) I've successfully installed the preempt patches on my old machine - P2 400 and on my new XP1800+ Athlon - all seems to work well. One other note that I'm not sure if I mentioned in my original e-mail - I always compile linux+xfs using kgcc. I'm tossing around using a newer gcc sometime soon, however, kgcc has always worked very well for me. Good luck, -Walt Gabe E. Nydick wrote: > Eric, > > Is the preempt patch supposed to fix anything with athlons? > > Thanks, > --------------------- > Gabe E. Nydick > Systems Lead > ClubPhoto, Inc. > P: 408.423.6611 > F: 408.557.6799 > ---------------------- > > On 12/7/01 10:56 AM, "Eric Sandeen" wrote: > > >>On Fri, 2001-12-07 at 12:48, Gabe E. Nydick wrote: >> >>>I looked yesterday for the 2.4.16 patches, couldn't find them, and I >>>couldn't find the preempt patches either, can you point me to them? >>> >>ftp://oss.sgi.com/projects/xfs/download/patches/2.4.16/ >> >>preempt patches: >>http://www.tech9.net/rml/linux/ >> >>-Eric >> > From owner-linux-xfs@oss.sgi.com Fri Dec 7 22:00:19 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB860J801590 for linux-xfs-outgoing; Fri, 7 Dec 2001 22:00:19 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB860Co01563 for ; Fri, 7 Dec 2001 22:00:12 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via SMTP id VAA05585 for ; Fri, 7 Dec 2001 21:00:07 -0800 (PST) mail_from (nathans@wobbly.melbourne.sgi.com) Received: from wobbly.melbourne.sgi.com (wobbly.melbourne.sgi.com [134.14.55.135]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id PAA27310; Sat, 8 Dec 2001 15:58:44 +1100 Received: (from nathans@localhost) by wobbly.melbourne.sgi.com (SGI-8.9.3/8.9.3) id PAA56284; Sat, 8 Dec 2001 15:58:42 +1100 (AEDT) Date: Sat, 8 Dec 2001 15:58:41 +1100 From: Nathan Scott To: "Stephen C . Tweedie" , Andreas Gruenbacher Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [PATCH] Revised extended attributes interface Message-ID: <20011208155841.A56289@wobbly.melbourne.sgi.com> References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <20011207202036.J2274@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20011207202036.J2274@redhat.com>; from sct@redhat.com on Fri, Dec 07, 2001 at 08:20:36PM +0000 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Fri, Dec 07, 2001 at 08:20:36PM +0000, Stephen C. Tweedie wrote: > Hi, > hi Stephen, > This is looking OK as far as EAs go. However, there is still no > mention of ACLs specifically, except an oblique reference to > "system.posix_acl_access". Yup - there's little mention of ACLs because they are only an optional, higher-level consumer of the API, & so didn't seem appropriate to document here. We have implemented POSIX ACLs above this interface - there is source to new versions of Andreas' user tools here: http://oss.sgi.com/cgi-bin/cvsweb.cgi/linux-2.4-xfs/cmd/acl2 These have been tested with XFS and seem to work fine, so we are ready to transition over from our old implementation to this new one. In a way there's consensus wrt how to do POSIX ACLs on Linux now, as both the ext2/ext3 and XFS ACL projects will be using the same tools, libraries, etc. In terms of other ACL types, I don't know of anyone actively working on any. The existence of a POSIX ACL implementation using attributes system.posix_acl_access and system.posix_acl_default doesn't preclude other types of ACLs from being implemented (obviously using different attributes) as well of course, if someone had an itch to scratch. cheers. -- Nathan From owner-linux-xfs@oss.sgi.com Sat Dec 8 00:05:42 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB885g603447 for linux-xfs-outgoing; Sat, 8 Dec 2001 00:05:42 -0800 Received: from walden.phpwebhosting.com (walden.phpwebhosting.com [64.65.61.214]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB885Yo03423 for ; Sat, 8 Dec 2001 00:05:34 -0800 Received: (qmail 3570 invoked by uid 508); 8 Dec 2001 07:05:23 -0000 Received: from unknown (HELO localhost) (24.196.64.71) by walden.phpwebhosting.com with SMTP; 8 Dec 2001 07:05:23 -0000 Date: Sat, 8 Dec 2001 01:06:18 -0600 Subject: Re: Linux w/XFS on Athlon systems Content-Type: text/plain; charset=US-ASCII; format=flowed Mime-Version: 1.0 (Apple Message framework v475) From: Ben Gollmer To: linux-xfs@oss.sgi.com Content-Transfer-Encoding: 7bit In-Reply-To: <1007664581.7654.61.camel@squash.mw.mediaone.net> Message-Id: <13A17680-EBAA-11D5-B4A1-003065B71D96@jatosoft.com> X-Mailer: Apple Mail (2.475) Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Running RH 7.2 with XFS 1.0.2 on an Athlon 1 GHz + Abit KT7E mainboard here (3 IBM 45 GB IDE HD's installed). Works great. Does installing a 'canned' distro with XFS work? The problem may be something in your kernel compilation. Ben On Thursday, December 6, 2001, at 12:49 PM, Joe Landman wrote: > Drew: > > I would go with a distribution that supports XFS. That would be > Mandrake 8.1 or MSC.Linux from MSC.Software (I work the them). Both > work quite well on Athlon hardware (Tyan, Abit, Asus, Gigabyte, ECS). > RedHat 7.2 can be shoehorned to take XFS as can be seen from the > excellent work of the folks at SGI who redid the RedHat installer. It > is a shame the folks over at RedHat havent seen the light. > > Do the standard diagnostic stuff on your hardware. AMD Athlon + > Motherboard + Linux + XFS does work quite well. Only problems I have had > seems to have stemmed from devfs. A devfs=nomount seems to make most of > them go away. > > Joe > > On Thu, 2001-12-06 at 13:07, Drew 'Cheetah!' Maxwell wrote: >> Fellow XFS users: >> -------------------------------------------------------------------------------- >> My company has been working with XFS-Linux on Intel systems for some >> time >> and has had some difficulties getting a stable Linux w/XFS release to >> work >> on some of our older K6-2 systems. Instead of butting our heads against >> a >> case of potentially suspect hardware, we're looking to upgrade to a >> group >> of Athlon-based systems. In this, I have a few questions: >> >> 1. What can those of you running Linux w/XFS recommend for motherboards >> to >> go with an Athlon processor? >> (We might as well get hardware we know works by experience.) >> 2. Are they any special changes to your kernels used to work correctly >> with >> an Athlon-based system? >> >> Our initial attempts at doing an Athlon + Linux + XFS solution haven't >> worked. >> We've tried the following methods: >> 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using >> gcc >> 2.4.14 source from kernel.org, patched to XFS 1.0.2, compiled using >> kgcc >> Pre-patched XFS 1.0.2 kernel source from SGI, compiled using kgcc >> >> Any assistance is greatly appreciated. >> >> --Drew 'Cheetah!' Maxwell [cheetah@clubphoto.com] >> ClubPhoto (http://www.clubphoto.com) Pictures, Prints, Gifts, and >> More! > > > From owner-linux-xfs@oss.sgi.com Sat Dec 8 13:18:47 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB8LIlu18705 for linux-xfs-outgoing; Sat, 8 Dec 2001 13:18:47 -0800 Received: from thebsh.namesys.com (thebsh.namesys.com [212.16.0.238]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB8LIfo18682 for ; Sat, 8 Dec 2001 13:18:42 -0800 Received: (qmail 24989 invoked from network); 8 Dec 2001 20:18:34 -0000 Received: from unknown (HELO namesys.com) (212.16.7.95) by thebsh.namesys.com with SMTP; 8 Dec 2001 20:18:34 -0000 Message-ID: <3C127551.90305@namesys.com> Date: Sat, 08 Dec 2001 23:17:21 +0300 From: Hans Reiser User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.6) Gecko/20011120 X-Accept-Language: en-us MIME-Version: 1.0 To: Nathan Scott CC: "Stephen C . Tweedie" , Andreas Gruenbacher , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [PATCH] Revised extended attributes interface References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <20011207202036.J2274@redhat.com> <20011208155841.A56289@wobbly.melbourne.sgi.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Nathan Scott wrote: > > >In a way there's consensus wrt how to do POSIX ACLs on Linux >now, as both the ext2/ext3 and XFS ACL projects will be using >the same tools, libraries, etc. In terms of other ACL types, >I don't know of anyone actively working on any. > > We are taking a very different approach to EAs (and thus to ACLs) as described in brief at www.namesys.com/v4/v4.html. We don't expect anyone to take us seriously on it before it works, but silence while coding does not equal consensus.;-) In essence, we think that if a file can't do what an EA can do, then you need to make files able to do more. It is very important not to reduce the amount of closure (as in mathematical closure) within the namespace, and creating EAs that cannot be accessed as files reduces closure. The same argument applies to streams, but it is kind of interesting to see people argue against streams for this reason, and then embrace EAs. Kind of leaves you wondering whether their hatred of streams was really any deeper than streams aren't what they are used to from Unix. Hans From owner-linux-xfs@oss.sgi.com Sat Dec 8 13:44:00 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB8Li0C19126 for linux-xfs-outgoing; Sat, 8 Dec 2001 13:44:00 -0800 Received: from zeus.city.tvnet.hu (zeus.city.tvnet.hu [195.38.100.182]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB8Lhvo19103 for ; Sat, 8 Dec 2001 13:43:57 -0800 Received: (from sferi@localhost) by zeus.city.tvnet.hu (8.11.6/8.11.6) id fB8KhHG03499; Sat, 8 Dec 2001 21:43:17 +0100 X-Authentication-Warning: zeus.city.tvnet.hu: sferi set sender to sferi@dumballah.tvnet.hu using -f Subject: redhat 7.2 xfs network installHi! From: Sipos Ferenc To: linux-xfs@oss.sgi.com Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 08 Dec 2001 21:43:16 +0100 Message-Id: <1007844196.1308.0.camel@zeus.city.tvnet.hu> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi! Does anybody know of a network installable redhat 7.2 xfs system on a fast european server? I'm not used to write cds for years. Thx. Paco From owner-linux-xfs@oss.sgi.com Sat Dec 8 16:13:25 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB90DP021714 for linux-xfs-outgoing; Sat, 8 Dec 2001 16:13:25 -0800 Received: from c007.snv.cp.net (c007-h008.c007.snv.cp.net [209.228.33.214]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB90DKo21690 for ; Sat, 8 Dec 2001 16:13:20 -0800 Received: (cpmta 25761 invoked from network); 8 Dec 2001 15:13:13 -0800 Received: from 64.34.34.69 (HELO lightning) by smtp.telocity.com (209.228.33.214) with SMTP; 8 Dec 2001 15:13:13 -0800 X-Sent: 8 Dec 2001 23:13:13 GMT From: "Matt Avila" To: Subject: FW: Suggestions on Emulex Driver build w XFS Date: Sat, 8 Dec 2001 18:19:07 -0500 Message-ID: <004101c1803e$bc897fe0$c8010a0a@lightning> MIME-Version: 1.0 X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.2627 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Importance: Normal Content-Disposition: inline Content-Type: text/plain Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk > All, > > I've been following this discussion group for some time and am > currently running kernel 2.4.9-13 w/ XFS 1.0.2 on an Intel board/CPU > combo. I've also installed dual Emulex LP8000's that will be AL_PA > connected to storage in the interim. Ill be moving these to FC_SW in > the near future. > > My questions are I have a few choices in building the driver (for the > fibre cards) either as a standalone module or to install it under the > /usr/src/linux kernel subtree. Any suggestions on "best practices" for > configuring this with XFS? Anyway, I've got a bit more than 1/2 TB > that I'm just itching to format & run with. > > Thanks in advance > > Matt [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Sat Dec 8 17:49:55 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB91nta24416 for linux-xfs-outgoing; Sat, 8 Dec 2001 17:49:55 -0800 Received: from smtp-server1.tampabay.rr.com (smtp-server1.tampabay.rr.com [65.32.1.34]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB91nmo24394 for ; Sat, 8 Dec 2001 17:49:48 -0800 Received: from ieee.org (653425hfc157.tampabay.rr.com [65.34.25.157]) by smtp-server1.tampabay.rr.com (8.11.2/8.11.2) with ESMTP id fB90nhO08419; Sat, 8 Dec 2001 19:49:43 -0500 (EST) Message-ID: <3C12B48A.ADF7F134@ieee.org> Date: Sat, 08 Dec 2001 19:47:06 -0500 From: Bryan-TheBS-Smith Organization: SmithConcepts, Inc. X-Mailer: Mozilla 4.78 [en] (X11; U; Linux 2.4.9-13SGI_XFS_1.0.2smp i686) X-Accept-Language: en MIME-Version: 1.0 To: Sipos Ferenc CC: linux-xfs@oss.sgi.com Subject: Re: redhat 7.2 xfs network installHi! References: <1007844196.1308.0.camel@zeus.city.tvnet.hu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Sipos Ferenc wrote: > Does anybody know of a network installable redhat 7.2 xfs > system on a fast european server? I'm not used to write cds > for years. Thx. Er, are you unfamilar with how to do NFS installs? Or don't you have the space to store it? You can download either the individual files or the CD (and mount on the loopback device), put them in a centralized area, NFS export that area, and boot from the "bootnet" floppy to connect to that NFS server. Just FYI. Let me know if you can't do that. -- TheBS -- Bryan "TheBS" Smith mailto:b.j.smith@ieee.org chat:thebs413 Engineer AbsoluteValue Systems, Inc. http://www.linux-wlan.org President SmithConcepts, Inc. http://www.SmithConcepts.com ---------------------------------------------------------------- "in a situation where you have someone against you ... an ac- cusation ... can lead to time in jail. This fear can be more effective in controlling a group of people than the enforcement of the law as is. -- Joao Miguel Neves on the US DMCA and EU CD From owner-linux-xfs@oss.sgi.com Sat Dec 8 18:33:47 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB92Xlc25235 for linux-xfs-outgoing; Sat, 8 Dec 2001 18:33:47 -0800 Received: from mail.ocs.com.au (mail.ocs.com.au [203.34.97.2]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB92Xfo25213 for ; Sat, 8 Dec 2001 18:33:42 -0800 Received: (qmail 31307 invoked from network); 9 Dec 2001 01:33:39 -0000 Received: from ocs3.intra.ocs.com.au (192.168.255.3) by mail.ocs.com.au with SMTP; 9 Dec 2001 01:33:39 -0000 Received: by ocs3.intra.ocs.com.au (Postfix, from userid 16331) id 8C0AC300090; Sun, 9 Dec 2001 12:33:37 +1100 (EST) Received: from ocs3.intra.ocs.com.au (localhost [127.0.0.1]) by ocs3.intra.ocs.com.au (Postfix) with ESMTP id 36ABA96; Sun, 9 Dec 2001 12:33:37 +1100 (EST) X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 From: Keith Owens To: "Matt Avila" Cc: linux-xfs@oss.sgi.com Subject: Re: FW: Suggestions on Emulex Driver build w XFS In-reply-to: Your message of "Sat, 08 Dec 2001 18:19:07 CDT." <004101c1803e$bc897fe0$c8010a0a@lightning> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 09 Dec 2001 12:33:31 +1100 Message-ID: <17308.1007861611@ocs3.intra.ocs.com.au> Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Sat, 8 Dec 2001 18:19:07 -0500, "Matt Avila" wrote: > I've been following this discussion group for some time and am > currently running kernel 2.4.9-13 w/ XFS 1.0.2 on an Intel board/CPU > combo. I've also installed dual Emulex LP8000's that will be AL_PA > connected to storage in the interim. Ill be moving these to FC_SW in > the near future. > > My questions are I have a few choices in building the driver (for the > fibre cards) either as a standalone module or to install it under the > /usr/src/linux kernel subtree. Any suggestions on "best practices" for > configuring this with XFS? Anyway, I've got a bit more than 1/2 TB > that I'm just itching to format & run with. Speaking for myself, not for SGI. IMHO building the Emulex drivers into the kernel violates the GPL. lpfc-i386.tar contains a binary only object (lpfcdriver) which is copied to lpfcdriver.o then linked into the kernel. I discussed this with Linus when Emulex first asked for help on linking their drivers into the kernel, his response was unambiguous, binary only objects linked into the kernel violate the GPL. From owner-linux-xfs@oss.sgi.com Sat Dec 8 19:07:46 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB937k525789 for linux-xfs-outgoing; Sat, 8 Dec 2001 19:07:46 -0800 Received: from c007.snv.cp.net (c007-h014.c007.snv.cp.net [209.228.33.221]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB937do25767 for ; Sat, 8 Dec 2001 19:07:39 -0800 Received: (cpmta 18114 invoked from network); 8 Dec 2001 18:07:32 -0800 Received: from 64.34.34.69 (HELO lightning) by smtp.telocity.com (209.228.33.221) with SMTP; 8 Dec 2001 18:07:32 -0800 X-Sent: 9 Dec 2001 02:07:32 GMT From: "Matt Avila" To: "'Keith Owens'" Cc: Subject: RE: FW: Suggestions on Emulex Driver build w XFS Date: Sat, 8 Dec 2001 21:13:27 -0500 Message-ID: <005501c18057$173cb890$c8010a0a@lightning> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.2627 In-Reply-To: <17308.1007861611@ocs3.intra.ocs.com.au> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Importance: Normal Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Keith, Thanks for your input.... I guess I understand the "legalese" that your referring to. I'm most questioning that in the Emulex readme there are references to two makefiles, one for a standalone and the other for the kernel, both of which are part of the Emulex TARball. They too are vague (basically nonexistent) on which is the preferred way to build the driver. I'm more adapt at setting up large storage installations and tuning accordingly, not at tweaking at the kernel level, so I was hoping to hear more about the technical implications of building into the kernel subtree. I do not want to get into a situation that if/when the system is further upgraded that I have to worry about continual kernel modifications. Would I be correct in assuming that <> I build and install as loadable modules this will not be an issue? Thanks again for your input Matt -----Original Message----- From: owner-linux-xfs@oss.sgi.com [mailto:owner-linux-xfs@oss.sgi.com] On Behalf Of Keith Owens Sent: Saturday, December 08, 2001 8:34 PM To: Matt Avila Cc: linux-xfs@oss.sgi.com Subject: Re: FW: Suggestions on Emulex Driver build w XFS On Sat, 8 Dec 2001 18:19:07 -0500, "Matt Avila" wrote: > I've been following this discussion group for some time and am > currently running kernel 2.4.9-13 w/ XFS 1.0.2 on an Intel board/CPU > combo. I've also installed dual Emulex LP8000's that will be AL_PA > connected to storage in the interim. Ill be moving these to FC_SW in > the near future. > > My questions are I have a few choices in building the driver (for the > fibre cards) either as a standalone module or to install it under the > /usr/src/linux kernel subtree. Any suggestions on "best practices" for > configuring this with XFS? Anyway, I've got a bit more than 1/2 TB > that I'm just itching to format & run with. Speaking for myself, not for SGI. IMHO building the Emulex drivers into the kernel violates the GPL. lpfc-i386.tar contains a binary only object (lpfcdriver) which is copied to lpfcdriver.o then linked into the kernel. I discussed this with Linus when Emulex first asked for help on linking their drivers into the kernel, his response was unambiguous, binary only objects linked into the kernel violate the GPL. From owner-linux-xfs@oss.sgi.com Sat Dec 8 19:22:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB93M3Z26158 for linux-xfs-outgoing; Sat, 8 Dec 2001 19:22:03 -0800 Received: from mail.ocs.com.au (mail.ocs.com.au [203.34.97.2]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB93Lxo26135 for ; Sat, 8 Dec 2001 19:21:59 -0800 Received: (qmail 31760 invoked from network); 9 Dec 2001 02:21:56 -0000 Received: from ocs3.intra.ocs.com.au (192.168.255.3) by mail.ocs.com.au with SMTP; 9 Dec 2001 02:21:56 -0000 Received: by ocs3.intra.ocs.com.au (Postfix, from userid 16331) id CA137300090; Sun, 9 Dec 2001 13:21:55 +1100 (EST) Received: from ocs3.intra.ocs.com.au (localhost [127.0.0.1]) by ocs3.intra.ocs.com.au (Postfix) with ESMTP id 82A5896; Sun, 9 Dec 2001 13:21:55 +1100 (EST) X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 From: Keith Owens To: "Matt Avila" Cc: linux-xfs@oss.sgi.com Subject: Re: FW: Suggestions on Emulex Driver build w XFS In-reply-to: Your message of "Sat, 08 Dec 2001 21:13:27 CDT." <005501c18057$173cb890$c8010a0a@lightning> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 09 Dec 2001 13:21:50 +1100 Message-ID: <17853.1007864510@ocs3.intra.ocs.com.au> Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Sat, 8 Dec 2001 21:13:27 -0500, "Matt Avila" wrote: >I do not want to get into a situation that if/when the system is further >upgraded that I have to worry about continual kernel modifications. >Would I be correct in assuming that <> I build and install as >loadable modules this will not be an issue? That is a different question. Binary only modules are accepted by Linus, but that just covers the legal angle. It says nothing about whether the drivers will work with a new kernel or not, you have to ask the driver supplier if it is safe to upgrade your kernel. The Emulex web site says that it was tested on RH 7.1 and kernel 2.4.4, for any other kernel version you will have to ask Emulex. From owner-linux-xfs@oss.sgi.com Sun Dec 9 05:56:14 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB9DuET06316 for linux-xfs-outgoing; Sun, 9 Dec 2001 05:56:14 -0800 Received: from smtpzilla2.xs4all.nl (smtpzilla2.xs4all.nl [194.109.127.138]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB9Du6o06284 for ; Sun, 9 Dec 2001 05:56:06 -0800 Received: from auto-nb1.xs4all.nl (qn-212-58-167-191.quicknet.nl [212.58.167.191]) by smtpzilla2.xs4all.nl (8.12.0/8.12.0) with ESMTP id fB9Ctt4Y028164; Sun, 9 Dec 2001 13:55:56 +0100 (CET) Message-Id: <4.3.2.7.2.20011209134615.02a845a8@pop.xs4all.nl> X-Sender: knuffie@pop.xs4all.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Sun, 09 Dec 2001 13:52:26 +0100 To: "Matt Avila" , From: Seth Mos Subject: Re: FW: Suggestions on Emulex Driver build w XFS In-Reply-To: <004101c1803e$bc897fe0$c8010a0a@lightning> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 18:19 8-12-2001 -0500, Matt Avila wrote: > > All, > > > > I've been following this discussion group for some time and am > > currently running kernel 2.4.9-13 w/ XFS 1.0.2 on an Intel board/CPU > > combo. I've also installed dual Emulex LP8000's that will be AL_PA > > connected to storage in the interim. Ill be moving these to FC_SW in > > the near future. > > > > My questions are I have a few choices in building the driver (for the > > fibre cards) either as a standalone module or to install it under the > > /usr/src/linux kernel subtree. Any suggestions on "best practices" for > > configuring this with XFS? Anyway, I've got a bit more than 1/2 TB > > that I'm just itching to format & run with. Binary drivers are dangerous waters and upgrading your kernel may suddenly break the driver or the kernel and produce dataloss, lockups and corruption. Looking at what adaptec did with their raid solution which is found onboard of a lot of prefab servers and goes by the name of aacraid. Better avoid that. People have seen a myriad of problems in the past and even today when some boxes still die under heavy load. Alan Cox has just made a real linux driver for these cards which after some testing by people shows that even at 0.9.2 the card survives all forms of crashtesting which is something incredible for a complete rewrite. I have no word on the performance yet. We officially scrapped these cards 1 month after purchase of the machine and used a card with a driver that was generally available and working. I suspect that if you go the closed binary path it will be full of bumps and potholes. A risk that I am not willing to take. Cheers -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Sun Dec 9 07:00:17 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fB9F0HT07560 for linux-xfs-outgoing; Sun, 9 Dec 2001 07:00:17 -0800 Received: from ente.berdmann.de (m-dialin-916.addcom.de [62.96.172.204]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fB9F0Fo07537 for ; Sun, 9 Dec 2001 07:00:15 -0800 Received: from apollo.berdmann.de ([192.168.1.2] helo=berdmann.de) by ente.berdmann.de with esmtp (Exim 3.22 #1) id 16D4Uu-0000wi-00 for linux-xfs@oss.sgi.com; Sun, 09 Dec 2001 15:00:04 +0100 Message-ID: <3C136E64.F8D54E83@berdmann.de> Date: Sun, 09 Dec 2001 15:00:04 +0100 From: "Bernhard R. Erdmann" X-Mailer: Mozilla 4.78 [en] (X11; U; Linux 2.4.14-xfs i586) X-Accept-Language: de, en, fr MIME-Version: 1.0 To: Linux XFS Mailing List Subject: xfsrestore forgets to set owner/perm on . Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk When transferring a filesystem with "xfsdump -J /bla | xfsrestore - .", xfsrestore does not set owner nor permissions of /bla on . From owner-linux-xfs@oss.sgi.com Sun Dec 9 16:39:17 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBA0dHl29082 for linux-xfs-outgoing; Sun, 9 Dec 2001 16:39:17 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBA0dDo29060 for ; Sun, 9 Dec 2001 16:39:13 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via SMTP id PAB03669 for ; Sun, 9 Dec 2001 15:39:10 -0800 (PST) mail_from (ivanr@sgi.com) Received: from omen.melbourne.sgi.com (omen.melbourne.sgi.com [134.14.55.139]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id KAA04811; Mon, 10 Dec 2001 10:37:54 +1100 From: ivanr@sgi.com Received: from localhost (ivanr@localhost) by omen.melbourne.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id KAA91139; Mon, 10 Dec 2001 10:37:52 +1100 (EST) X-Authentication-Warning: omen.melbourne.sgi.com: ivanr owned process doing -bs Date: Mon, 10 Dec 2001 10:37:52 +1100 X-X-Sender: ivanr@omen.melbourne.sgi.com To: "Bernhard R. Erdmann" cc: Linux XFS Mailing List Subject: Re: xfsrestore forgets to set owner/perm on . In-Reply-To: <3C136E64.F8D54E83@berdmann.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Sun, 9 Dec 2001, Bernhard R. Erdmann wrote: > When transferring a filesystem with "xfsdump -J /bla | xfsrestore - .", > xfsrestore does not set owner nor permissions of /bla on . xfsrestore does not create its destination directory, it simply restores the dump to the directory specified. I don't think it is up to xfsrestore to change the ownership or permissions of the owning directory -- it can't assume what other uses that directory might have. Eg. "xfsdump - /filesystem | xfsrestore - /tmp" Ivan -- Ivan Rayner ivanr@sgi.com From owner-linux-xfs@oss.sgi.com Sun Dec 9 17:32:34 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBA1WYC30660 for linux-xfs-outgoing; Sun, 9 Dec 2001 17:32:34 -0800 Received: from c007.snv.cp.net (c007-h013.c007.snv.cp.net [209.228.33.220]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBA1WPo30637 for ; Sun, 9 Dec 2001 17:32:25 -0800 Received: (cpmta 2960 invoked from network); 9 Dec 2001 16:32:18 -0800 Received: from 64.34.34.69 (HELO lightning) by smtp.telocity.com (209.228.33.220) with SMTP; 9 Dec 2001 16:32:18 -0800 X-Sent: 10 Dec 2001 00:32:18 GMT From: "Matt Avila" To: "'Seth Mos'" , Subject: RE: FW: Suggestions on Emulex Driver build w XFS Date: Sun, 9 Dec 2001 19:38:24 -0500 Message-ID: <000001c18112$fa949340$c8010a0a@lightning> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.2627 In-Reply-To: <4.3.2.7.2.20011209134615.02a845a8@pop.xs4all.nl> X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Importance: Normal Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Seth, Many thanks for your input on this. Ill go the non-kernel-mod route. Any Idea if Alan has his drivers available? Id be inclined to give them a go as opposed to compiling the Emulex drivers from what you have mentioned. This box is not a production machine but rater a system at my home that I'm using for some testing and a repository for a few files laying round so I'm really not afraid to try things with it. At any rate, thanks for your input. Matt -----Original Message----- From: owner-linux-xfs@oss.sgi.com [mailto:owner-linux-xfs@oss.sgi.com] On Behalf Of Seth Mos Sent: Sunday, December 09, 2001 7:52 AM To: Matt Avila; linux-xfs@oss.sgi.com Subject: Re: FW: Suggestions on Emulex Driver build w XFS At 18:19 8-12-2001 -0500, Matt Avila wrote: > > All, > > > > I've been following this discussion group for some time and am > > currently running kernel 2.4.9-13 w/ XFS 1.0.2 on an Intel board/CPU > > combo. I've also installed dual Emulex LP8000's that will be AL_PA > > connected to storage in the interim. Ill be moving these to FC_SW in > > the near future. > > > > My questions are I have a few choices in building the driver (for > > the fibre cards) either as a standalone module or to install it > > under the /usr/src/linux kernel subtree. Any suggestions on "best > > practices" for configuring this with XFS? Anyway, I've got a bit > > more than 1/2 TB that I'm just itching to format & run with. Binary drivers are dangerous waters and upgrading your kernel may suddenly break the driver or the kernel and produce dataloss, lockups and corruption. Looking at what adaptec did with their raid solution which is found onboard of a lot of prefab servers and goes by the name of aacraid. Better avoid that. People have seen a myriad of problems in the past and even today when some boxes still die under heavy load. Alan Cox has just made a real linux driver for these cards which after some testing by people shows that even at 0.9.2 the card survives all forms of crashtesting which is something incredible for a complete rewrite. I have no word on the performance yet. We officially scrapped these cards 1 month after purchase of the machine and used a card with a driver that was generally available and working. I suspect that if you go the closed binary path it will be full of bumps and potholes. A risk that I am not willing to take. Cheers -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Sun Dec 9 22:41:41 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBA6ffR11773 for linux-xfs-outgoing; Sun, 9 Dec 2001 22:41:41 -0800 Received: from sgi.com (sgi.SGI.COM [192.48.153.1]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBA6fYo11748 for ; Sun, 9 Dec 2001 22:41:34 -0800 Received: from boing.melbourne.sgi.com (boing.melbourne.sgi.com [134.14.55.141]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id VAB08628 for ; Sun, 9 Dec 2001 21:41:25 -0800 (PST) mail_from (tes@boing.melbourne.sgi.com) Received: (from tes@localhost) by boing.melbourne.sgi.com (SGI-8.9.3/8.9.3) id QAA59381; Mon, 10 Dec 2001 16:39:58 +1100 (AEDT) Date: Mon, 10 Dec 2001 16:39:57 +1100 From: Timothy Shimmin To: Juergen Hasch Cc: ivanr@sgi.com, linux-xfs@oss.sgi.com Subject: Re: Again: xfsrestore assertion failure Message-ID: <20011210163957.Q61575@boing.melbourne.sgi.com> References: <169t5H-0FoJM0C@fwd03.sul.t-online.com> <16AFxD-1LpiRkC@fwd06.sul.t-online.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0us In-Reply-To: <16AFxD-1LpiRkC@fwd06.sul.t-online.com>; from Hasch@t-online.de on Sat, Dec 01, 2001 at 08:37:27PM +0100 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi Juergen, On Sat, Dec 01, 2001 at 08:37:27PM +0100, Juergen Hasch wrote: > Am Friday 30 November 2001 20:12 schrieb Juergen Hasch: > > > In line 1508 of drive_scsitape.c I find the following values: > > first_mark_offset = 206e460000000000 > > file_offset = 400000 > > tape_recsz = 809efa800100000 > > Ivan, things go wrong in xfsdump. > When do_set_mark() sets first_mark_offset in the rec_hdr_t struct > it doesn't endian convert it. It is also not endian converted later > The following patch solves this, however you could also do the > endian conversion later in do_write(). > ...Juergen > > --- cmd/xfsdump/common/drive_scsitape.c.orig Fri Nov 30 19:43:48 2001 > +++ cmd/xfsdump/common/drive_scsitape.c Sat Dec 1 20:25:35 2001 > @@ -1940,7 +1940,7 @@ > rechdrp = ( rec_hdr_t * )contextp->dc_recp; > if ( rechdrp->first_mark_offset == -1LL ) { > ASSERT( nextoff != -1LL ); > - rechdrp->first_mark_offset = nextoff; > + rechdrp->first_mark_offset = INT_GET(nextoff,ARCH_CONVERT); > } > > /* put the mark on the tail of the queue. > > Thanks muchly for the patch, Juergen, - I see the problem. However, I've changed the code so that all the endian-conversion for the rechdr is done in the one place just prior to the write call (in write_record()). (My fix seems a bit clearer and safer.) I'll check it into our src tree as soon as it becomes unlocked (from some internal cloning which is going on). Cheers, Tim. From owner-linux-xfs@oss.sgi.com Mon Dec 10 02:23:41 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAANf125315 for linux-xfs-outgoing; Mon, 10 Dec 2001 02:23:41 -0800 Received: from smtpzilla1.xs4all.nl (smtpzilla1.xs4all.nl [194.109.127.137]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAANao25293 for ; Mon, 10 Dec 2001 02:23:37 -0800 Received: from auto-nb1.xs4all.nl (coltex.xs4all.nl [213.84.127.168]) by smtpzilla1.xs4all.nl (8.12.0/8.12.0) with ESMTP id fBA9NTbh033142; Mon, 10 Dec 2001 10:23:30 +0100 (CET) Message-Id: <4.3.2.7.2.20011210094023.02c0bb20@pop.xs4all.nl> X-Sender: knuffie@pop.xs4all.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Mon, 10 Dec 2001 10:20:02 +0100 To: "Matt Avila" , From: Seth Mos Subject: RE: FW: Suggestions on Emulex Driver build w XFS In-Reply-To: <000001c18112$fa949340$c8010a0a@lightning> References: <4.3.2.7.2.20011209134615.02a845a8@pop.xs4all.nl> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk At 19:38 9-12-2001 -0500, Matt Avila wrote: >Seth, > >Many thanks for your input on this. Ill go the non-kernel-mod route. Any >Idea if Alan has his drivers available? Id be inclined to give them a go >as opposed to compiling the Emulex drivers from what you have mentioned. The aacraid drivers are for the Adaptec raid controllers which are u160 scsi raid controllers. It's not for fiberchannel drives. Although the Perc3/QC is a fibrechannel device but I don't know what chip it uses. Some of the cards are Megaraid, others are adaptec or Qlogic I believe. You can find the aacraid driver on the ftp.linux.org.uk machine or look in the archives of the aacraid-devel list. >This box is not a production machine but rater a system at my home that >I'm using for some testing and a repository for a few files laying round >so I'm really not afraid to try things with it. Testing is good. -- Seth Every program has two purposes one for which it was written and another for which it wasn't I use the last kind. From owner-linux-xfs@oss.sgi.com Mon Dec 10 04:52:45 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBACqjs27410 for linux-xfs-outgoing; Mon, 10 Dec 2001 04:52:45 -0800 Received: from sisko.scot.redhat.com (pc-62-31-66-178-ed.blueyonder.co.uk [62.31.66.178]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBACqWo27381 for ; Mon, 10 Dec 2001 04:52:33 -0800 Received: (from sct@localhost) by sisko.scot.redhat.com (8.11.6/8.11.2) id fBABq9I06340; Mon, 10 Dec 2001 11:52:09 GMT Date: Mon, 10 Dec 2001 11:52:09 +0000 From: "Stephen C. Tweedie" To: Nathan Scott Cc: "Stephen C . Tweedie" , Andreas Gruenbacher , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [PATCH] Revised extended attributes interface Message-ID: <20011210115209.C1919@redhat.com> References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <20011207202036.J2274@redhat.com> <20011208155841.A56289@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20011208155841.A56289@wobbly.melbourne.sgi.com>; from nathans@sgi.com on Sat, Dec 08, 2001 at 03:58:41PM +1100 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, On Sat, Dec 08, 2001 at 03:58:41PM +1100, Nathan Scott wrote: > On Fri, Dec 07, 2001 at 08:20:36PM +0000, Stephen C. Tweedie wrote: > > > This is looking OK as far as EAs go. However, there is still no > > mention of ACLs specifically, except an oblique reference to > > "system.posix_acl_access". > > Yup - there's little mention of ACLs because they are only an > optional, higher-level consumer of the API, & so didn't seem > appropriate to document here. Unfortunately, if there are many filesystems wanting to use posix ACLs, then standardising the API is still desirable. > We have implemented POSIX ACLs above this interface - there > is source to new versions of Andreas' user tools here: > http://oss.sgi.com/cgi-bin/cvsweb.cgi/linux-2.4-xfs/cmd/acl2 > These have been tested with XFS and seem to work fine, so we > are ready to transition over from our old implementation to > this new one. But the ACL encoding is still hobbled: there's no namespace for credentials other than uid/gid. This has been brought up before, but it's worth going over some of the things we'd like to be able to do with extended credentials again: * NFSv4. NFSv4 credentials are of the form "user@realm", and an NFSv4 server needs to be able to apply ACLs using such credentials so that it can securely serve users in foreign realms. * Kerberos single-signon. I want to be able to get a kerberos login ticket on the desktop in front of me and access files in my entire organisation securely. I want to be able to login to remote systems in different departments and still have ACLs work. So "foo@SALES.CO.COM" might login to a machine in the "DEVEL.CO.COM", and would only get a "guest" uid, but the ACL system would allow access based on the full "foo@SALES.CO.COM" credentials. * Samba. Is there any reason not to allow an NT SID to be used as the credential for an ACL? * Sub-IDs. There was a beautiful paper presented at a recent Usenix in which the concept of user-manageable sub-ids was presented. I am on a secure intranet, but I'm constantly accessing untrusted data. Every time Mozilla accesses a web site I am potentially vulnerable to web rendering bugs which could allow a site to take over my machine. Plugins such as flash just make the matter worse. Even in the home environment we'd like to make it easy to allow multiuser games to be run without compromising the whole local system. The sub-id concept proposes allowing users to create process groups with restricted rights to the system. I would _really_ like to give Mozilla write access to ~/tmp and ~/.mozilla, but not to the rest of my homedir. Can't I use a "sct/mozilla" credential for my ACLs? Authentication is about *much* more than just local uid/gids, but the current EA/ACL specs are creating an implicit standard for ACLs without addressing any of these concerns. > The existence of a POSIX ACL implementation using attributes > system.posix_acl_access and system.posix_acl_default doesn't > preclude other types of ACLs from being implemented (obviously > using different attributes) as well of course, if someone had > an itch to scratch. I am not talking about other types of ACLs! I am talking about *POSIX* ACLs, but using a credentials namespace which is more than just uid/gid. Only the credentials change: the rest of the POSIX semantics still apply. The CITI NFSv4 implementation is already doing POSIX ACLs and GSSAPI krb5 authentication on top of the bestbits API, so we already have at least one application ready and waiting to use such an extension. Cheers, Stephen From owner-linux-xfs@oss.sgi.com Mon Dec 10 07:33:19 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAFXJm29617 for linux-xfs-outgoing; Mon, 10 Dec 2001 07:33:19 -0800 Received: from otto.plogic.internal (plogic.com [209.92.41.135]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAFXGo29595 for ; Mon, 10 Dec 2001 07:33:16 -0800 Received: from rlatham by otto.plogic.internal with local (Exim 3.33 #1 (Debian)) id 16DRUY-00012t-00 for ; Mon, 10 Dec 2001 09:33:14 -0500 Date: Mon, 10 Dec 2001 09:33:14 -0500 From: Rob Latham To: linux-xfs@oss.sgi.com Subject: Re: FW: Suggestions on Emulex Driver build w XFS Message-ID: <20011210093314.X28778@otto.plogic.internal> References: <004101c1803e$bc897fe0$c8010a0a@lightning> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <004101c1803e$bc897fe0$c8010a0a@lightning>; from mavila@telocity.com on Sat, Dec 08, 2001 at 06:19:07PM -0500 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Sat, Dec 08, 2001 at 06:19:07PM -0500, Matt Avila wrote: > > My questions are I have a few choices in building the driver (for the > > fibre cards) either as a standalone module or to install it under the > > /usr/src/linux kernel subtree. Any suggestions on "best practices" for > > configuring this with XFS? Anyway, I've got a bit more than 1/2 TB > > that I'm just itching to format & run with. build it as a standalone module. it'll be easier to upgrade ( only need to take storage off line ), easier to test and easier to tune, ( though the default parameters have so far been fine for me ) ==rob -- [ Rob Latham Developer, Admin, Alchemist ] [ Paralogic Inc. - www.plogic.com ] [ ] [ EAE8 DE90 85BB 526F 3181 1FCF 51C4 B6CB 08CC 0897 ] From owner-linux-xfs@oss.sgi.com Mon Dec 10 08:00:40 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAG0ew30644 for linux-xfs-outgoing; Mon, 10 Dec 2001 08:00:40 -0800 Received: from bear.monarch.net (bear.monarch.net [24.244.0.67]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAG0bo30622 for ; Mon, 10 Dec 2001 08:00:37 -0800 Received: (qmail 25267 invoked from network); 10 Dec 2001 15:00:28 -0000 Received: from ca-mnet244-12-43.monarch.net (HELO lustre.cfs) (24.244.12.43) by bear.monarch.net with SMTP; 10 Dec 2001 15:00:28 -0000 Received: by lustre.cfs (Postfix, from userid 500) id 7B1B56C28A; Mon, 10 Dec 2001 08:00:03 -0700 (MST) Date: Mon, 10 Dec 2001 08:00:03 -0700 From: "Peter J. Braam" To: "Stephen C. Tweedie" Cc: Nathan Scott , Andreas Gruenbacher , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [PATCH] Revised extended attributes interface Message-ID: <20011210080002.A25961@lustre.cfs> References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <20011207202036.J2274@redhat.com> <20011208155841.A56289@wobbly.melbourne.sgi.com> <20011210115209.C1919@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20011210115209.C1919@redhat.com>; from sct@redhat.com on Mon, Dec 10, 2001 at 11:52:09AM +0000 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Mon, Dec 10, 2001 at 11:52:09AM +0000, Stephen C. Tweedie wrote: > * Sub-IDs. > > There was a beautiful paper presented at a recent Usenix in which the > concept of user-manageable sub-ids was presented. Stephen, Do you have a ref for that? Thanks! From owner-linux-xfs@oss.sgi.com Mon Dec 10 08:05:15 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAG5FF30856 for linux-xfs-outgoing; Mon, 10 Dec 2001 08:05:15 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAG5Ao30834 for ; Mon, 10 Dec 2001 08:05:10 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id QAA1868168 for ; Mon, 10 Dec 2001 16:05:07 +0100 (CET) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id JAA3817327; Mon, 10 Dec 2001 09:03:50 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id JAA54099; Mon, 10 Dec 2001 09:03:49 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fBAF2eT27184; Mon, 10 Dec 2001 09:02:40 -0600 Subject: Re: [XFS-1.0.2,LVM-1.0.1,linux-2.4.14] "mount -U " does not work with XFS. From: Steve Lord To: "Karl M. Hegbloom" Cc: linux-xfs@oss.sgi.com In-Reply-To: <871yi6zq3z.fsf@juniper.intra.microsharp.com> References: <871yi6zq3z.fsf@juniper.intra.microsharp.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 10 Dec 2001 09:02:40 -0600 Message-Id: <1007996560.27133.2.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Fri, 2001-12-07 at 15:18, Karl M. Hegbloom wrote: > [ I am not subscribed to the list; Please Cc. ] > > I'm using "mount" version 2.11m-1 (debian "sid"), Linux 2.4.14 from > Debian "kernel-source", version 2.4.14-1, patched with Debian > supplied "kernel-patch-xfs", version 1.0.2-2, and LVM release 1.0.1 > patched in by myself from source obtained directly from the LVM > site... Since the mount by uuid code is in the mount command, it is possible you do not have a recent enough mount command there. Check for XFS specific code in the mount source. Steve > -- > mailto: (Karl M. Hegbloom) karlheg@microsharp.com > http://www.microsharp.com > phone://USA/WA/360-260-2066 > jabber: karlheg@jabber.org -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Mon Dec 10 08:10:51 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAGApd31093 for linux-xfs-outgoing; Mon, 10 Dec 2001 08:10:51 -0800 Received: from minnie.omroep.nl (minnie.omroep.nl [145.58.30.4]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAGAeo31068 for ; Mon, 10 Dec 2001 08:10:40 -0800 Received: (from smap@localhost) by minnie.omroep.nl (SGI-8.9.3/8.9.3) id QAA95527 for ; Mon, 10 Dec 2001 16:10:32 +0100 (CET) Received: from nos-smtp.nos.nl (145.58.12.125 "HELO nos-smtp.nos.nl") by minnie.omroep.nl with SMTP (smap v3.0 nederlandse publieke omroep) id xma5699828; Mon, 10 Dec 01 16:10:31 +0100 Received: FROM exchange.nos.nl BY nos-smtp.nos.nl ; Mon Dec 10 16:22:57 2001 +0100 Received: by EXCHANGE with Internet Mail Service (5.5.2650.21) id ; Mon, 10 Dec 2001 16:13:19 +0100 Message-ID: <35F87783B600D411A1430060943F469A589736@EXCHANGE> From: Matthijs van der Klip To: "'Dean Roehrich'" , Matthijs van der Klip Cc: Linux XFS Mailing List Subject: RE: using dmapi to sync filesystems (was using xfsdump...) Date: Mon, 10 Dec 2001 16:13:19 +0100 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Disposition: inline Content-Type: text/plain Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >Using XFS and the dmapi tests in cmd/xfstests/dmapi: > ># mount -o dmapi /dev/hda12 /mnts/dmi1 ># dm_create_session -i dean1 >ret=0 >newsid=1 ># set_disp -s 1 /mnts/dmi1 DM_EVENT_CREATE DM_EVENT_REMOVE ># set_eventlist -s 1 /mnts/dmi1 DM_EVENT_CREATE DM_EVENT_REMOVE ># touch /mnts/dmi1/foo1& >[1] 771 ># ># get_events 1 >rlenp is 96 >create: token=1 sequence=1 > parent dir 5a659482b0bbecee0e000000000000004000000000000000 > name foo1 > mode bits mode 100644: perm rw- r-- r--, type Regular File ># respond_event 1 1 1 0 ># >[1] + Done touch /mnts/dmi1/foo1& > >To get actual pathnames, well that's different. Our HSM doesn't need the >pathname, and so...it's not till now that I see maybe dm_handle_to_path() in >libdm is broken. I'm sure someone who needs it could come up with a fix >easily enough :) Hi Dean, Could you explain? I'm afraid I need some more guidance. Best regards, Matthijs van der Klip NOS - Dutch Public Broadcasting Organisation [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Mon Dec 10 08:18:31 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAGIVJ04681 for linux-xfs-outgoing; Mon, 10 Dec 2001 08:18:31 -0800 Received: from merlin.giref.ulaval.ca (postfix@merlin.giref.ulaval.ca [132.203.7.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAGIPo04657 for ; Mon, 10 Dec 2001 08:18:25 -0800 Received: from giref.ulaval.ca (castellane.giref.ulaval.ca [132.203.7.26]) by merlin.giref.ulaval.ca (Postfix) with ESMTP id 4AEC0DE47 for ; Mon, 10 Dec 2001 10:18:23 -0500 (EST) Message-ID: <3C14D23F.634B275@giref.ulaval.ca> Date: Mon, 10 Dec 2001 10:18:23 -0500 From: Luc Lalonde X-Mailer: Mozilla 4.79 [en] (X11; U; SunOS 5.8 sun4u) X-Accept-Language: en MIME-Version: 1.0 To: "linux-xfs@oss.sgi.com" Subject: Solaris clients causing NFS lockups Content-Type: multipart/mixed; boundary="------------54E66D03D9BC8CD79FFDB19D" Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk This is a multi-part message in MIME format. --------------54E66D03D9BC8CD79FFDB19D Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello Folks, I'm having trouble with lockups with my server running linux-2.4.14-xfs. My server does a complete lockup under heavy access load from Solaris NFS clients. I reboot the machine and if locks up within a few minutes. The only way to get things going again is to shutdown all my Solaris clients, restart my Linux NFS server and then proceed restarting the SUN client boxes. This could be completely off topic...I'm just wondering if anyone has seen the same problem? Perhaps is a matter of setting some mount options that are specific to Solaris8 clients. Cheers. -- Luc Lalonde, Responsable du reseau GIREF Telephone: (418) 656-2131 poste 6623 Courriel: llalonde@giref.ulaval.ca --------------54E66D03D9BC8CD79FFDB19D Content-Type: text/x-vcard; charset=us-ascii; name="llalonde.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Luc Lalonde Content-Disposition: attachment; filename="llalonde.vcf" begin:vcard n:Lalonde;Luc x-mozilla-html:FALSE org:Universite Laval;GIREF adr:;;;;;; version:2.1 email;internet:llalonde@giref.ulaval.ca title:Administateur de reseau x-mozilla-cpt:;0 fn:Luc Lalonde end:vcard --------------54E66D03D9BC8CD79FFDB19D-- From owner-linux-xfs@oss.sgi.com Mon Dec 10 08:39:40 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAGdeR05759 for linux-xfs-outgoing; Mon, 10 Dec 2001 08:39:40 -0800 Received: from rj.sgi.com (rj.SGI.COM [204.94.215.100]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAGdbo05736 for ; Mon, 10 Dec 2001 08:39:37 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by rj.sgi.com (8.11.4/8.11.4/linux-outbound_gateway-1.1) with ESMTP id fBAFdUY31398 for ; Mon, 10 Dec 2001 07:39:30 -0800 Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id JAA3821073 for ; Mon, 10 Dec 2001 09:38:15 -0600 (CST) Received: from slobber.americas.sgi.com (slobber.americas.sgi.com [128.162.187.52]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id JAA53144 for ; Mon, 10 Dec 2001 09:38:14 -0600 (CST) Received: from slobber.americas.sgi.com by slobber.americas.sgi.com (SGI-8.9.3/SGI-client-1.7) via ESMTP id JAA05140; Mon, 10 Dec 2001 09:38:14 -0600 (CST) Message-Id: <200112101538.JAA05140@slobber.americas.sgi.com> To: Linux XFS Mailing List Subject: Re: using dmapi to sync filesystems (was using xfsdump...) Date: Mon, 10 Dec 2001 09:38:13 -0600 From: Dean Roehrich Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >From: Matthijs van der Klip >Hi Dean, > >Could you explain? I'm afraid I need some more guidance. Is there some particular piece of my note you want explained, or did you want me to go over all of it again, in more detail? Dean From owner-linux-xfs@oss.sgi.com Mon Dec 10 08:53:04 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAGr4I08138 for linux-xfs-outgoing; Mon, 10 Dec 2001 08:53:04 -0800 Received: from mx2.elte.hu (mx2.elte.hu [157.181.151.9]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAGqxo08116 for ; Mon, 10 Dec 2001 08:52:59 -0800 Received: from chiara.elte.hu (chiara.elte.hu [157.181.150.200]) by mx2.elte.hu (Postfix) with ESMTP id DDB444800A for ; Mon, 10 Dec 2001 16:52:51 +0100 (CET) Received: by chiara.elte.hu (Postfix, from userid 17000) id C59701FCE; Mon, 10 Dec 2001 16:52:50 +0100 (CET) Date: Mon, 10 Dec 2001 16:52:50 +0100 From: KELEMEN Peter To: linux-xfs@oss.sgi.com Subject: 2.4.16-xfs oopses on mount (highmem?) Message-ID: <20011210165249.A32478@chiara.elte.hu> Reply-To: KELEMEN Peter Mail-Followup-To: linux-xfs@oss.sgi.com Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit User-Agent: Mutt/1.3.23i Organization: ELTE Eotvos Lorand University of Sciences, Budapest, Hungary X-GPG-KeyID: 1024D/EE4C26E8 2000-03-20 X-GPG-Fingerprint: D402 4AF3 7488 165B CC34 4147 7F0C D922 EE4C 26E8 X-PGP-KeyID: 1024/45F83E45 1998/04/04 X-PGP-Fingerprint: 26 87 63 4B 07 28 1F AD 6D AA B5 8A D6 03 0F BF X-Comment: Personal opinion. Paragraphs might have been reformatted. X-Copyright: Forwarding or publishing without permission is prohibited. X-Accept-Language: hu,en Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hello, Our dual PPro 200 oopses on filesystem mount with 2.4.16-xfs from CVS as of last Saturday. The machine has 1GB memory, the stack trace was: sys_write()->linvfs_write()->xfs_write()->pagebuf_generic_file_write() ->_pagebuf_file_write()->_pagebuf_do_delwri()->kunmap_high() Sorry for the terse trace, I had no serial console at hand. The filesystem to be mounted is a 2G root filesystem on RAID1. Peter PS: kgcc-compiled as always. -- .+'''+. .+'''+. .+'''+. .+'''+. .+'' Kelemen Péter / \ / \ / fuji@elte.hu .+' `+...+' `+...+' `+...+' `+...+' From owner-linux-xfs@oss.sgi.com Mon Dec 10 08:57:01 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAGv1F08354 for linux-xfs-outgoing; Mon, 10 Dec 2001 08:57:01 -0800 Received: from sisko.scot.redhat.com (pc-62-31-66-178-ed.blueyonder.co.uk [62.31.66.178]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAGuto08328 for ; Mon, 10 Dec 2001 08:56:55 -0800 Received: (from sct@localhost) by sisko.scot.redhat.com (8.11.6/8.11.2) id fBAFuS210842; Mon, 10 Dec 2001 15:56:28 GMT Date: Mon, 10 Dec 2001 15:56:28 +0000 From: "Stephen C. Tweedie" To: "Peter J. Braam" Cc: "Stephen C. Tweedie" , Nathan Scott , Andreas Gruenbacher , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [PATCH] Revised extended attributes interface Message-ID: <20011210155628.E1919@redhat.com> References: <20011205143209.C44610@wobbly.melbourne.sgi.com> <20011207202036.J2274@redhat.com> <20011208155841.A56289@wobbly.melbourne.sgi.com> <20011210115209.C1919@redhat.com> <20011210080002.A25961@lustre.cfs> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20011210080002.A25961@lustre.cfs>; from braam@clusterfs.com on Mon, Dec 10, 2001 at 08:00:03AM -0700 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, On Mon, Dec 10, 2001 at 08:00:03AM -0700, Peter J. Braam wrote: > On Mon, Dec 10, 2001 at 11:52:09AM +0000, Stephen C. Tweedie wrote: > > > * Sub-IDs. > > > > There was a beautiful paper presented at a recent Usenix in which the > > concept of user-manageable sub-ids was presented. > > Stephen, Do you have a ref for that? http://www.usenix.org/publications/library/proceedings/usenix01/freenix01/ioannidis.html Cheers, Stephen From owner-linux-xfs@oss.sgi.com Mon Dec 10 09:01:19 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAH1JW08609 for linux-xfs-outgoing; Mon, 10 Dec 2001 09:01:19 -0800 Received: from filesrv1.baby-dragons.com (ns1.baby-dragons.com [199.33.245.254]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAH1Fo08587 for ; Mon, 10 Dec 2001 09:01:15 -0800 Received: from localhost (babydr@localhost) by filesrv1.baby-dragons.com (8.11.4/8.11.4) with ESMTP id fBAG06C07447; Mon, 10 Dec 2001 11:00:06 -0500 Date: Mon, 10 Dec 2001 11:00:06 -0500 (EST) From: "Mr. James W. Laferriere" To: "Stephen C. Tweedie" cc: "Peter J. Braam" , Nathan Scott , Andreas Gruenbacher , , , Subject: Re: [PATCH] Revised extended attributes interface In-Reply-To: <20011210155628.E1919@redhat.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hello Stephen , Is this the only attribution ? Just love those 'we won't share security info with you unless you are member or pay.' . Sorry , JimL On Mon, 10 Dec 2001, Stephen C. Tweedie wrote: > Hi, > > On Mon, Dec 10, 2001 at 08:00:03AM -0700, Peter J. Braam wrote: > > On Mon, Dec 10, 2001 at 11:52:09AM +0000, Stephen C. Tweedie wrote: > > > > > * Sub-IDs. > > > > > > There was a beautiful paper presented at a recent Usenix in which the > > > concept of user-manageable sub-ids was presented. > > > > Stephen, Do you have a ref for that? > > http://www.usenix.org/publications/library/proceedings/usenix01/freenix01/ioannidis.html > > Cheers, > Stephen > - > To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > +------------------------------------------------------------------+ | James W. Laferriere | System Techniques | Give me VMS | | Network Engineer | P.O. Box 854 | Give me Linux | | babydr@baby-dragons.com | Coudersport PA 16915 | only on AXP | +------------------------------------------------------------------+ From owner-linux-xfs@oss.sgi.com Mon Dec 10 09:02:15 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAH2Fb08743 for linux-xfs-outgoing; Mon, 10 Dec 2001 09:02:15 -0800 Received: from pooh.lsc.hu (IDENT:postfix@lsc.net23.hu [195.56.172.131]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAH22o08716 for ; Mon, 10 Dec 2001 09:02:02 -0800 Received: by pooh.lsc.hu (Postfix, from userid 547) id 09C06E057D; Mon, 10 Dec 2001 17:22:22 +0100 (CET) Date: Mon, 10 Dec 2001 17:22:21 +0100 From: GCS To: linux-xfs@oss.sgi.com Subject: Re: 2.4.16-xfs oopses on mount (highmem?) Message-ID: <20011210172221.A11888@lsc.hu> References: <20011210165249.A32478@chiara.elte.hu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20011210165249.A32478@chiara.elte.hu>; from fuji@elte.hu on Mon, Dec 10, 2001 at 04:52:50PM +0100 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi Peter, On Mon, Dec 10, 2001 at 04:52:50PM +0100, KELEMEN Peter wrote: > Our dual PPro 200 oopses on filesystem mount with 2.4.16-xfs from > CVS as of last Saturday. The machine has 1GB memory, the stack > trace was: Hmmm. It can be a problem with the standard kernel even. You may know Bra, the admin of ftp.fsn.hu . He also told that with a very similar config one of his machines dies if he uses the 2.4.x series. Can you test it without XFS support? Also, as I see the upcoming 2.4.17 will contain several small bugfixes, maybe one of it will help you out. Anyway, when SGI would like to update the 2.4.x tree to the pre kernel? Cheers, GCS -- Your mouse has moved. Windows NT must be restarted for the change to take effect. Reboot now? [OK] From owner-linux-xfs@oss.sgi.com Mon Dec 10 09:04:20 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAH4KN08903 for linux-xfs-outgoing; Mon, 10 Dec 2001 09:04:20 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAH43o08872 for ; Mon, 10 Dec 2001 09:04:03 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id RAA1929878 for ; Mon, 10 Dec 2001 17:03:59 +0100 (CET) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id KAA3812519; Mon, 10 Dec 2001 10:02:42 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id KAA74592; Mon, 10 Dec 2001 10:02:42 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fBAG1WP31062; Mon, 10 Dec 2001 10:01:32 -0600 Subject: Re: 2.4.16-xfs oopses on mount (highmem?) From: Steve Lord To: KELEMEN Peter Cc: linux-xfs@oss.sgi.com In-Reply-To: <20011210165249.A32478@chiara.elte.hu> References: <20011210165249.A32478@chiara.elte.hu> Content-Type: multipart/mixed; boundary="=-H3g3Gas+savtdoNQ41zc" X-Mailer: Evolution/1.0 (Preview Release) Date: 10 Dec 2001 10:01:32 -0600 Message-Id: <1008000092.28858.11.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk --=-H3g3Gas+savtdoNQ41zc Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable On Mon, 2001-12-10 at 09:52, KELEMEN Peter wrote: > Hello, >=20 > Our dual PPro 200 oopses on filesystem mount with 2.4.16-xfs from > CVS as of last Saturday. The machine has 1GB memory, the stack > trace was: >=20 > sys_write()->linvfs_write()->xfs_write()->pagebuf_generic_file_write() > ->_pagebuf_file_write()->_pagebuf_do_delwri()->kunmap_high() >=20 > Sorry for the terse trace, I had no serial console at hand. The > filesystem to be mounted is a 2G root filesystem on RAID1. Probably my fault, some changes went into page_buf_io.c late last week, if you back off one version on this file it should work again. Some internal admin activity at the moment means I cannot update the cvs trees, you could also try the attached patch which builds, but has not been tested. Steve >=20 > Peter >=20 > PS: kgcc-compiled as always. >=20 > --=20 > .+'''+. .+'''+. .+'''+. .+'''+. .+'' > Kelemen P=E9ter / \ / \ / fuji@elte.hu > .+' `+...+' `+...+' `+...+' `+...+' --=20 Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com --=-H3g3Gas+savtdoNQ41zc Content-Disposition: attachment; filename=pagebuf.patch Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=ISO-8859-1 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Index: linux/fs/pagebuf/page_buf_io.c =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D --- /usr/tmp/TmpDir.31049-0/linux/fs/pagebuf/page_buf_io.c_1.101 Mon Dec 10= 10:02:06 2001 +++ linux/fs/pagebuf/page_buf_io.c Mon Dec 10 09:59:36 2001 @@ -1031,6 +1031,8 @@ =20 err =3D __copy_from_user(kaddr + offset, buf, bytes); if (err) { + err =3D -EFAULT; + kunmap(page); ClearPageUptodate(page); goto unlock; } @@ -1043,7 +1045,6 @@ buf +=3D bytes;=20 =20 unlock: - kunmap(page); UnlockPage(page); page_cache_release(page); if (err < 0) @@ -1269,6 +1270,7 @@ =20 if (status) { status =3D -EFAULT; + kunmap(page); ClearPageUptodate(page); goto unlock; } @@ -1281,7 +1283,6 @@ written +=3D bytes; =20 unlock: - kunmap(page); SetPageReferenced(page); UnlockPage(page); page_cache_release(page); --=-H3g3Gas+savtdoNQ41zc-- From owner-linux-xfs@oss.sgi.com Mon Dec 10 09:05:50 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAH5o509128 for linux-xfs-outgoing; Mon, 10 Dec 2001 09:05:50 -0800 Received: from minnie.omroep.nl (minnie.omroep.nl [145.58.30.4]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAH5lo09104 for ; Mon, 10 Dec 2001 09:05:47 -0800 Received: (from smap@localhost) by minnie.omroep.nl (SGI-8.9.3/8.9.3) id RAA81902 for ; Mon, 10 Dec 2001 17:05:40 +0100 (CET) Received: from nos-smtp.nos.nl (145.58.12.125 "HELO nos-smtp.nos.nl") by minnie.omroep.nl with SMTP (smap v3.0 nederlandse publieke omroep) id xma5241868; Mon, 10 Dec 01 17:05:38 +0100 Received: FROM exchange.nos.nl BY nos-smtp.nos.nl ; Mon Dec 10 17:18:05 2001 +0100 Received: by EXCHANGE with Internet Mail Service (5.5.2650.21) id ; Mon, 10 Dec 2001 17:08:28 +0100 Message-ID: <35F87783B600D411A1430060943F469A589739@EXCHANGE> From: Matthijs van der Klip To: Linux XFS Mailing List Subject: RE: using dmapi to sync filesystems (was using xfsdump...) Date: Mon, 10 Dec 2001 17:08:27 +0100 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Disposition: inline Content-Type: text/plain Content-Transfer-Encoding: 7bit Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >Is there some particular piece of my note you want explained, or did you want >me to go over all of it again, in more detail? Taken this discussion off the list suspecting it might be too detailed... :) Best regards, Matthijs van der Klip [[HTML alternate version deleted]] From owner-linux-xfs@oss.sgi.com Mon Dec 10 09:07:35 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAH7Z209269 for linux-xfs-outgoing; Mon, 10 Dec 2001 09:07:35 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAH7Uo09244 for ; Mon, 10 Dec 2001 09:07:30 -0800 Received: from zeus-fddi.americas.sgi.com (zeus-fddi.americas.sgi.com [128.162.8.103]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id RAA1854029 for ; Mon, 10 Dec 2001 17:07:27 +0100 (CET) mail_from (lord@sgi.com) Received: from daisy-e185.americas.sgi.com (daisy-e185.americas.sgi.com [128.162.185.214]) by zeus-fddi.americas.sgi.com (8.9.3/americas-smart-nospam1.1) with ESMTP id KAA3812179; Mon, 10 Dec 2001 10:06:10 -0600 (CST) Received: from jen.americas.sgi.com (jen.americas.sgi.com [128.162.187.49]) by daisy-e185.americas.sgi.com (SGI-8.9.3/SGI-server-1.7) with ESMTP id KAA75819; Mon, 10 Dec 2001 10:06:10 -0600 (CST) Received: by jen.americas.sgi.com (8.11.6/SGI-client-1.7) id fBAG50731075; Mon, 10 Dec 2001 10:05:00 -0600 Subject: Re: 2.4.16-xfs oopses on mount (highmem?) From: Steve Lord To: GCS Cc: linux-xfs@oss.sgi.com In-Reply-To: <20011210172221.A11888@lsc.hu> References: <20011210165249.A32478@chiara.elte.hu> <20011210172221.A11888@lsc.hu> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0 (Preview Release) Date: 10 Dec 2001 10:05:00 -0600 Message-Id: <1008000300.28858.14.camel@jen.americas.sgi.com> Mime-Version: 1.0 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk On Mon, 2001-12-10 at 10:22, GCS wrote: > Hi Peter, > > On Mon, Dec 10, 2001 at 04:52:50PM +0100, KELEMEN Peter wrote: > > Our dual PPro 200 oopses on filesystem mount with 2.4.16-xfs from > > CVS as of last Saturday. The machine has 1GB memory, the stack > > trace was: > Hmmm. It can be a problem with the standard kernel even. You may know Bra, > the admin of ftp.fsn.hu . He also told that with a very similar config one > of his machines dies if he uses the 2.4.x series. Can you test it without XFS > support? > Also, as I see the upcoming 2.4.17 will contain several small bugfixes, maybe > one of it will help you out. Anyway, when SGI would like to update the 2.4.x > tree to the pre kernel? We are no longer going to be updating the 2.4 tree to anything but actual 2.4 releases. Given the fact that we are starting to track 2.5 as well as 2.4 we needed to stop doing something else. Steve > > Cheers, GCS > -- > Your mouse has moved. > Windows NT must be restarted for the change to take effect. > Reboot now? [OK] -- Steve Lord voice: +1-651-683-3511 Principal Engineer, Filesystem Software email: lord@sgi.com From owner-linux-xfs@oss.sgi.com Mon Dec 10 09:19:29 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAHJTo09610 for linux-xfs-outgoing; Mon, 10 Dec 2001 09:19:29 -0800 Received: from sisko.scot.redhat.com (pc-62-31-66-178-ed.blueyonder.co.uk [62.31.66.178]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAHJKo09588 for ; Mon, 10 Dec 2001 09:19:20 -0800 Received: (from sct@localhost) by sisko.scot.redhat.com (8.11.6/8.11.2) id fBAGFXR10899; Mon, 10 Dec 2001 16:15:33 GMT Date: Mon, 10 Dec 2001 16:15:33 +0000 From: "Stephen C. Tweedie" To: "Mr. James W. Laferriere" Cc: "Stephen C. Tweedie" , "Peter J. Braam" , Nathan Scott , Andreas Gruenbacher , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-xfs@oss.sgi.com Subject: Re: [PATCH] Revised extended attributes interface Message-ID: <20011210161533.G1919@redhat.com> References: <20011210155628.E1919@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from babydr@baby-dragons.com on Mon, Dec 10, 2001 at 11:00:06AM -0500 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, On Mon, Dec 10, 2001 at 11:00:06AM -0500, Mr. James W. Laferriere wrote: > > Hello Stephen , Is this the only attribution ? > Just love those 'we won't share security info with you unless you > are member or pay.' . Sorry , JimL There are other references in the paper: I've appended them below. One, in particular, seems to talk about quite similar concepts: http://www.usenix.org/publications/library/proceedings/sec2000/acharya.html Cheers, Stephen [1] NJS JavaScript Interpreter. http://www.bbassett.net/njs/. [2] The OpenBSD Operating System. http://www.openbsd.org/. [3] World Wide Web Consortium. http://www.w3.org/. [4] Anurag Acharya and Mandar Raje. Map- box: Using parameterized behavior classes to confine applications. In Proceedings of the 2000 USENIX Security Symposium, pages 1-17, Denver, CO, August 2000. [5] Andrew Berman, Virgil Bourassa, and Erik Selberg. TRON: Process-Specific File Protection for the UNIX Operating System. In USENIX 1995 Technical Conference, New Orleans, Louisiana, January 1995. [6] David Flanagan. JavaScript The De nitive Guide. O'Reilly, 1998. [7] Tim Fraser, Lee Badger, and Mark Feldman. Hardening COTS Software with Generic Software Wrappers. In Proceedings of the IEEE Symposium on Security and Privacy, Oakland, CA, May 1999. [8] Ian Goldberg, David Wagner, Randi Thomas, and Eric A. Brewer. A Secure Environment for Untrusted Helper Applications. In USENIX 1996 Technical Conference, 1996. [9] Li Gong. Inside Java 2 Platform Security. Addison-Wesley, 1999. [10] James Gosling, Bill Joy, and Guy Steele. The Java Language Specification. Addison Wesley, Reading, 1996. [11] http://www.cert.org/advisories/. [12] Sotiris Ioannidis and Steven M. Bellovin. Sub-Operating Systems: A New Approach to Application Security. Technical Report MS-CIS-01- 06, University of Pennsylvania, February 2000. [13] R. Kaplan. SUID and SGID Based Attacks on UNIX: a Look at One Form of then Use and Abuse of Privileges. Computer Security Journal, 9(1):73-7, 1993. [14] Jacob Y. Levy, Laurent Demailly, John K. Ousterhout, and Brent B. Welch. The Safe-Tcl Security Model. In USENIX 1998 Annual Technical Conference, New Orleans, Louisiana, June 1998. [15] Gary McGraw and Edward W. Felten. Java Security: hostile applets, holes and antidotes. Wiley, New York, NY, 1997. [16] G. C. Necula and P. Lee. Safe, Untrusted Agents using Proof-Carrying Code. In Lecture Notes in Computer Science Special Issue on Mobile Agents, October 1997. [17] Dan S. Wallach, Dirk Balfanz, Drew Dean, and Edward W. Felten. Extensible Security Architectures for Java. In Proceedings of the 16th ACM Symposium on Operating Systems Principles, October 1997. From owner-linux-xfs@oss.sgi.com Mon Dec 10 09:33:26 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAHXQI09892 for linux-xfs-outgoing; Mon, 10 Dec 2001 09:33:26 -0800 Received: from mail.loewe-komp.de (mail.loewe-komp.de [62.156.155.230]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAHXJo09870 for ; Mon, 10 Dec 2001 09:33:19 -0800 Received: from loewe-komp.de (pippin [192.168.169.19]) by mail.loewe-komp.de (8.11.0/8.11.0/SuSE Linux 8.11.0-0.4) with ESMTP id fBAGZJS12606; Mon, 10 Dec 2001 17:35:20 +0100 X-Authentication-Warning: mail.loewe-komp.de: Host pippin [192.168.169.19] claimed to be loewe-komp.de Message-ID: <3C14E449.612ACFB0@loewe-komp.de> Date: Mon, 10 Dec 2001 17:35:21 +0100 From: Peter =?iso-8859-1?Q?W=E4chtler?= Organization: LOEWE. Hannover X-Mailer: Mozilla 4.78 [en] (X11; U; Linux 2.4.16 i686) X-Accept-Language: de, en MIME-Version: 1.0 To: Luc Lalonde CC: "linux-xfs@oss.sgi.com" Subject: Re: Solaris clients causing NFS lockups References: <3C14D23F.634B275@giref.ulaval.ca> Content-Type: text/plain; charset=iso-8859-1 X-MIME-Autoconverted: from 8bit to quoted-printable by mail.loewe-komp.de id fBAGZJS12606 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id fBAHXKo09871 Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Luc Lalonde wrote: > > I'm having trouble with lockups with my server running > linux-2.4.14-xfs. My server does a complete lockup under heavy access > load from Solaris NFS clients. I reboot the machine and if locks up > within a few minutes. The only way to get things going again is to > shutdown all my Solaris clients, restart my Linux NFS server and then > proceed restarting the SUN client boxes. > > This could be completely off topic...I'm just wondering if anyone has > seen the same problem? Perhaps is a matter of setting some mount > options that are specific to Solaris8 clients. > Well, I know the symptom with "locking up after server restart after crash". But no Solaris client involved - only Linux clients on Linux NFS server. I still wait for the next crash to happen (19 days uptime now). Normally it was crashing after ~14 days. I use now 2.4.9-xfs but will switch to 2.4.14-xfs if the crash occurs. I can't trigger the bug though (mounting our "compile environment" on 2 clients, compile some projects and pressing reset on the server) Instead of shutting down the clients, I pull the network cable, start in runlevel 1, wait, wait a bit longer, !"§$%&, plug in the cable and hey: it doesn't crash! From owner-linux-xfs@oss.sgi.com Mon Dec 10 10:40:22 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.2/8.11.3) id fBAIeMq11262 for linux-xfs-outgoing; Mon, 10 Dec 2001 10:40:22 -0800 Received: from microsharp.com (mail.microsharp.com [206.190.130.135]) by oss.sgi.com (8.11.2/8.11.3) with SMTP id fBAIeEo11240 for ; Mon, 10 Dec 2001 10:40:14 -0800 Received: from juniper.intra.microsharp.com (intra.microsharp.com [206.190.130.132] (may be forged)) by microsharp.com (8.11.2/8.11.2) with ESMTP id fBAHYiH15095; Mon, 10 Dec 2001 09:34:44 -0800 Received: from juniper.intra.microsharp.com (karlheg@localhost [127.0.0.1]) by juniper.intra.microsharp.com (8.12.1/8.12.1/Debian -2) with ESMTP id fBAHdvsX018302; Mon, 10 Dec 2001 09:39:57 -0800 Received: (from karlheg@localhost) by juniper.intra.microsharp.com (8.12.1/8.12.1/Debian -2) id fBAHdub4018298; Mon, 10 Dec 2001 09:39:56 -0800 To: Steve Lord Cc: linux-xfs@oss.sgi.com Subject: Re: [XFS-1.0.2,LVM-1.0.1,linux-2.4.14] "mount -U " does not work with XFS. References: <871yi6zq3z.fsf@juniper.intra.microsharp.com> <1007996560.27133.2.camel@jen.americas.sgi.com> X-Face: 1v=v?=]<7%cW&OL:Z"GCcdmIN&wp)w|1EKzyC6?qaX;55wwjPfKTm\'7x-gd+KEd`(o&(EQ%(B4]6sZ$|e_FBGTJOvJtHUM>flSNQqfdw.jYZ$jnCERoAX*{4lEJm?EB#V_<2\-#5%#Du/ExQR)juedX>!Cq#;UoAo?|Oo~BX@!fwIX From: karlheg@microsharp.com (Karl M. Hegbloom) In-Reply-To: <1007996560.27133.2.camel@jen.americas.sgi.com> Message-ID: <871yi3xbo4.fsf@juniper.intra.microsharp.com> Date: 10 Dec 2001 09:39:56 -0800 Lines: 69 User-Agent: Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.4 (Artificial Intelligence) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk >>>>> "Steve" == Steve Lord writes: Steve> On Fri, 2001-12-07 at 15:18, Karl M. Hegbloom wrote: >> [ I am not subscribed to the list; Please Cc. ] >> >> I'm using "mount" version 2.11m-1 (debian "sid"), Linux 2.4.14 from >> Debian "kernel-source", version 2.4.14-1, patched with Debian >> supplied "kernel-patch-xfs", version 1.0.2-2, and LVM release 1.0.1 >> patched in by myself from source obtained directly from the LVM >> site... Steve> Since the mount by uuid code is in the mount command, it is possible Steve> you do not have a recent enough mount command there. Check for XFS Steve> specific code in the mount source. I will look and see... I would assume that since the man page mentions XFS explicitly wrt UUID= mounting, that it is a recent enough version. % apt-get source mount Reading Package Lists... Done Building Dependency Tree... Done Need to get 1487kB of source archives. Get:1 http://http.us.debian.org unstable/main util-linux 2.11m-1 (dsc) [679B] Get:2 http://http.us.debian.org unstable/main util-linux 2.11m-1 (tar) [1437kB] Get:3 http://http.us.debian.org unstable/main util-linux 2.11m-1 (diff) [48.8kB] Fetched 1487kB in 10s (148kB/s) dpkg-source: extracting util-linux in util-linux-2.11m